pmcid
large_stringlengths 8
11
| pmid
large_stringlengths 0
8
| authors
large_stringlengths 0
17.7k
| title
large_stringlengths 7
1.37k
| publication_date
large_stringlengths 10
10
| keywords
large_stringlengths 0
5.03k
| abstract
large_stringlengths 0
21.7k
| text
large_stringlengths 282
5.89M
| high_quality
bool 1
class | medium_quality
bool 1
class | low_quality
bool 2
classes |
|---|---|---|---|---|---|---|---|---|---|---|
PMC9589879 | Walid F. Elkhatib,Shereen S. Abdelkareem,Wafaa S. Khalaf,Mona I. Shahin,Dounia Elfadil,Alaa Alhazmi,Ahmed I. El-Batal,Gharieb S. El-Sayyad | Narrative review on century of respiratory pandemics from Spanish flu to COVID-19 and impact of nanotechnology on COVID-19 diagnosis and immune system boosting | 24-10-2022 | SARS-CoV-2,Spanish flu,Nanotechnology,Immune system,Respiratory pandemics | The rise of the highly lethal severe acute respiratory syndrome-2 (SARS-2) as corona virus 2019 (COVID-19) reminded us of the history of other pandemics that happened in the last century (Spanish flu) and stayed in the current century, which include Severe-Acute-Respiratory-Syndrome (SARS), Middle-East-Respiratory-Syndrome (MERS), Corona Virus 2019 (COVID-19). We review in this report the newest findings and data on the origin of pandemic respiratory viral diseases, reservoirs, and transmission modes. We analyzed viral adaption needed for host switch and determinants of pathogenicity, causative factors of pandemic viruses, and symptoms and clinical manifestations. After that, we concluded the host factors associated with pandemics morbidity and mortality (immune responses and immunopathology, ages, and effect of pandemics on pregnancy). Additionally, we focused on the burdens of COVID-19, non-pharmaceutical interventions (quarantine, mass gatherings, facemasks, and hygiene), and medical interventions (antiviral therapies and vaccines). Finally, we investigated the nanotechnology between COVID-19 analysis and immune system boosting (Nanoparticles (NPs), antimicrobial NPs as antivirals and immune cytokines). This review presents insights about using nanomaterials to treat COVID-19, improve the bioavailability of the abused drugs, diminish their toxicity, and improve their performance. Graphical Abstract | Narrative review on century of respiratory pandemics from Spanish flu to COVID-19 and impact of nanotechnology on COVID-19 diagnosis and immune system boosting
The rise of the highly lethal severe acute respiratory syndrome-2 (SARS-2) as corona virus 2019 (COVID-19) reminded us of the history of other pandemics that happened in the last century (Spanish flu) and stayed in the current century, which include Severe-Acute-Respiratory-Syndrome (SARS), Middle-East-Respiratory-Syndrome (MERS), Corona Virus 2019 (COVID-19). We review in this report the newest findings and data on the origin of pandemic respiratory viral diseases, reservoirs, and transmission modes. We analyzed viral adaption needed for host switch and determinants of pathogenicity, causative factors of pandemic viruses, and symptoms and clinical manifestations. After that, we concluded the host factors associated with pandemics morbidity and mortality (immune responses and immunopathology, ages, and effect of pandemics on pregnancy). Additionally, we focused on the burdens of COVID-19, non-pharmaceutical interventions (quarantine, mass gatherings, facemasks, and hygiene), and medical interventions (antiviral therapies and vaccines). Finally, we investigated the nanotechnology between COVID-19 analysis and immune system boosting (Nanoparticles (NPs), antimicrobial NPs as antivirals and immune cytokines). This review presents insights about using nanomaterials to treat COVID-19, improve the bioavailability of the abused drugs, diminish their toxicity, and improve their performance.
The rapidly spreading throughout the existing highly contagious Severe-Acute-Respiratory-Syndrome-2 (SARS-2) or so-called Coronavirus 2019 (COVID-19) disease, reminded us of other pandemics that happened in the last century (H1N1 Spanish flu) and continued in the current century by (SARS, MERS, and COVID-19) [1–3]. After a series of respiratory viral pandemic diseases that started in 1918–1919 by a mysterious and fatal disease called H1N1 Spanish Flu and some researchers called it Mother of Pandemics [4], due to this pandemic has infected more than a third of the world's population and claimed an approximate 50 million lives, with untypically extreme clinical symptoms in young, formerly disease-free adults, the pandemic has been a major cause of death [5]. In this regard, SARS-CoV-2 and 1918 influenza A/H1N1 viruses have some common properties, such as being of similar basic reproduction numbers (R0), varying from 2 to 4, and similar shedding patterns from infectious patients, and hence likely to have similar generation gaps. In tandem, COVID-19 may have a latency period similar to that of influenza [6]. Then, in the current millennium, the world has witnessed persistent viral attacks from a novel viral family called Coronaviruses (CoVs) [7]. CoVs, containing an Orthocoronavirus subfamily and a Torovirinae subfamily. The Orthocoronavirinae subfamily comprises four genera: the alpha coronavirus, the beta coronavirus, the gamma coronavirus and the delta coronavirus [8]. Beta coronavirus genera encompass from Severe-Acute-Respiratory-Syndrome (SARS), Middle-East-Respiratory-Syndrome (MERS), human CoV-229E (HCoV-229E), HCoV-OC43, and Corona Virus-2019 (COVID-19) [3]. In 2002, 2012 and 2019, the world was attacked by three viral respiratory diseases by the SARS, the MERS and the COVID-19, accordingly. Corona viruses are enveloped, non-segmented, positive-sense, monostranded RNA viruses that show a characteristic appearance under negative-staining electron microscopy [9]. The WHO reported that in the southern Chinese province of Guangdong, on November 2002, no update was received from the Government of China up to the month's end of March, a massive 792 cases and 31 deaths were reported. The ministry of health of China reported more than 8,000 cases.It is estimated that there were 1 thousand cases of disease and about 774 deaths, giving a lethality rate of about 7%. The reservoir host of the infection is thought to be the Asian civet cat (Paguma larvata). It was expected that the host-to-human transmission sites would be open markets, as is the case with the current COVID-19 outbreak. [10]. The global SARS epidemic was successfully controlled in July 2003, and no cases of SARS have been reported since 2004. [11]. The emergence of SARS was followed by MERS as the second most important coronavirus causing a world-wide public health emergency. First appeared in Saudi Arabia (KSA) in 2012 when a patient aged 60 with severe pneumonia [10]. An epidemic of the virus only became apparent in 2014, with a total identified case count of 662 and a case fatality index of 32.97%. From 2014 to 2016, 1364 cases were observed in KSA [10]. Overall, 27 countries have been affected by MERS during the epidemics, in Asia, Europe, the Middle East, and North America. [12]. The cases identified beyond the Middle East, including the South Korea (SK) epidemic in where 186 people have been found to be infected as a result of a supraspreading, have involved transplant recipients who have already been infected in the Middle East. Since 2012, a total of 2494 lab-confirmed cases of MERS have been reported, with 858 associated deaths (case-fatality ratio of 34.4%) [10]. With regard to COVID-19, WHO has raised the risk level of the CoV crisis to "very high" on 28 February 2020. On 11 March, as COVID-19 incidents outside China have increased by a factor of 13 and the number of infected countries has increased threefold to over 118,000 registered cases in 114 different countries, with more than 4,000 fatalities, the WHO declared COVID-19 a global pandemic. Governments around the world are working to put countermeasures in place to hold the potentially deleterious effects. Health organizations are coordinating the flow of information and issuing directives to best minimize the impact of the pandemic threat. Meanwhile, researchers from all over the world are working intensively and information on the transmission mechanisms, the clinical spectrum of the new diagnosis of the disease, prevention and treatment approaches is developing rapidly. Many unknown factors concerning the virus-host dynamic and the progress of the epidemic remain, including the timing of its peak [13]. Nanomaterials (NMs) have specific features that are unique, which characterize them as outstanding materials able to apply in spectrum devices, sensors, and techniques used in virus’s detection, treatment, and virus’s elimination from the environment [14]. A major part of the applications of NMs is to predict and treat viruses in the health care and environment. In this review, we are presenting a brief about some of these uses.
The avian influenza is distinguished for its capacity to contaminate various animal diversity, including species such as bats, birds and mammals. Even though successful interspecies transmission is rare, it plays a pivotal role in the generation of new vectors of the pandemic [14]. In the pandemic of 1918–1919, March 1918 was the start of the spring wave. It spread across the Europe, US, and Asia over the next six months. Although levels of disease were high, mortality rates in most places were not significantly higher than usual. A second or autumn wave spread around the world from September to November 1918 and was very fatal. In many countries, a third wave occurred in early 1919. Contemporary observers concluded from the clinical similarities that they were seeing the same infection disease in successive waves [4]. The mild forms of the epidemics in all three waves were typical of the influenza seen in the 1889 pandemic and the avian flu epidemic and previous inter-pandemic years. In view of this, even the quick progression from uncomplicated influenza infection to fatal pneumonia, characteristic of the autumn and winter waves of 1918–1919, was observed in the few severe cases in the spring wave [4, 15]. Up until 2003, only 2 CoVs, Human CoV 229E (HCoV-229E) and HCoV-OC43 have been known to lead to human disease [16]. It manifests as mild symptoms such as a common cold in adults and more serious illness in infants, the old and immunocompromised people. In November 2002, numerous exceptional cases of "atypical pneumonia" of unexplained reason reported in the city of Foshan, Guangdong Province, In China, where many health staff have been contaminated [17]. This was introduced to Hong Kong on 21 February 2003 by a doctor who dealt with similar cases of SARS in the Chinese mainland, which resulted in widespread of serious pneumonia in Hong Kong and labelled by WHO as “severe-acute respiratory-syndrome” on March 15, 2003 [18, 19]. Months passed, and a number of incidents of SARS have been identified prior to SARS-CoV was identified. A new b-CoV (SARS-CoV) of lineage B was confirmed as the cause of the SARS pneumonia cases on 22 March 2003. The SARS-CoV pandemic has spread to 29 countries and regions. It was clear that the world's health, medical and scientific development communities were not sufficiently prepared for the emergence of SARS. [18]. Human-to-human chains of transmission have emerged in Canada, Toronto, Chinese Taipei, Hong Kong, China, Vietnam Singapore and Hanoi. The SARS epidemic had a short history and the WHO announced the winding down of the SARS epidemic in July 2003 [18]. Ten years since the last sign of SARS-CoV, in June 2012, a man died in KSA of serious pneumonia and kidney weakness [20]. A newly discovered corona-virus, the Middle-East-Respiratory-Syndrome-Coronavirus (MERS-CoV), has been identified from his sputum [21]. A group of severe-respiratory illness cases had emerged in April 2012 in Jordan in a hospital and were diagnosed in retrospect as MERS12, and a group of three MERS infected cases in the UK were detected in September 2012 MERS-CoV has continued to rise and expand outside the "Arabic peninsula”. As a consequence of travel by contaminated individuals; frequently these newly transmitted MERS cases have resulted in hospital-acquired transmission.[22]. In May 2015, the MERS outbreak in SK was triggered by a single returnee from the Middle East and affected sixteen clinics and 186 cases. As of 26 April 2016, 1,728 MERS cases have been confirmed, of which 624 fatalities in 27 different countries [22]. A pneumonia group of cases linked to a recently discovered β-coronavirus appeared in Wuhan, China, in December 2019. On January 12, 2020, the World Health Organization (WHO) designated this coronavirus as the 2019-novel coronavirus (2019-nCoV) (WHO). International Committee proposed naming the newly identified coronavirus as SARS-CoV-2, both reported on 11 February 2020. Chinese scientists rapidly isolated SARS-CoV-2 from a patient on 7 January 2020. They came out to genome sequencing of the SARS-CoV-2. As of 1 July 2021, 91,833 cases of COVID-19 have been confirmed in mainland China including 4636 deaths. The basic reproduction number “R0” of SARS-CoV-2 has been assessed by studies to be around 2.2 or more (range 1.4–6.5), and family pneumonia outbreak groupings add to the proof of a steadily growing COVID-19 epidemic through human-to-human transmission (Fig. 1) [23]. According to the WHO (https://covid19.who.int/); globally, as of 8:36 pm CEST, 14 April 2022, there have been 500,186,525 confirmed cases of COVID-19, including 6,190,349 deaths, reported to WHO. As of 18 April 2022, a total of 11,307,908,653 vaccine doses have been administered.
Influenza viruses in different species are known to originate in wild waterfowl. (Fig. 2) [24]. Whilst human-pig transmission subtypes have already been shown and substantiated, direct transmission between birds and humans has been less prevalent (as in the case of H9N2 and H5N1 subtypes) but has in some cases resulted in fatalities. [25]. Regarding the reservoir of coronaviridae family which consists of beta-coronavirus that include three pandemics recently (SARS, MERS, COVID-19). Bats are a huge natural reservoir wide range of CoVs, including SARS-CoV-like and MERS-CoV-like viruses (Fig. 2). Following the virus's genomic sequence, COVID-19 was studied, the genome for Bat CoV RaTG13 showed 96.2 percent overall identity of the genome series, indicating the CoV bat Even human SARS-CoV-2 may have the same parentage. In contrast, bats aren't accessible for purchase in this seafood marketplace business. By the way, synchronization of protein sequences and further phylogenetic study revealed similar residues receiver was found in several species, which offered more possibilities for alternate intermediate hosts; For starters, tortoises, pangolin, and snacks [23]. All pandemics respiratory viruses are zoonotic airborne RNA viruses that are rarely transmitted between native forms of humans but could mutate to make human transmission more effective. Frequent and approved transmission routes droppings (> 5 mm diameter, flying Transmission' < 1 m) make contact with the nose with viable viruses, mouth, eyes, or upper airway and 'airborne transmission' where droplets (5 mm diameter) are kernels [27]. 2002 and 2003 were examples of the H1N1 influenza pandemic in SARS and 2009. The function of 'direct communication of touch' (without the possibility of polluted surfaces) and 'indirect touch propagation' (including infected surfaces) of the distribution of such pandemic potential viruses have been controversial. Nonetheless, various reports and investigations have reported that indirect communication transmission is prevalent. The transmission path for other respiratory viruses, as well as influenza, under certain conditions [26].
Influenza virus replication occurs at the cellular level mainly in the epithelial cells of the intestinal tract in birds and in the epithelial cells of the respiratory tract in humans and other mammals [27]. In humans, ribonucleo-proteins (vRNPs) are subsequently transmitted into the nucleus of the diseased cells, in which viral RNA transcripts and replicates through the enzymatic activity of the viral polymerase complex attached to vRNPs [28]. The replication of viral RNA occurs via a positive intermediate, the complementary ribonucleoprotein complex [29]. Transcription of viral-RNA produces positive-stranded mRNA that is cap-linked and polyadenylated and then exported to the cytoplasm to be translated into viral proteins. [30]. Virus newly synthesised polymerases (PA, PB1,and PB2) and viral NP are imported into the nucleus to increase the rate of viral-RNA synthesis, while the viral membrane proteins HA, NA and M2 are transported and incorporated in the plasma membrane [31]. MERS-CoV and SARS-CoV have a specific coding mechanism in which about two thirds of the “viral RNA” is translated into two giant poly-proteins, while the remaining viral genome is transcribed in one nested series of subgenomic mRNAs [22]. Both pp1a and pp1ab, polyproteins encode sixteen non-structural proteins. (nsp1–nsp16) which bring up the viral replicase transcriptase complex [32]. The polyproteins are cleaved by papain-like protease (PLpro; corresponding to nsp3) two proteases, and The protease, 3C-like protease (3CLpro; corresponding to nsp5). nsps re-arrange membranes derivated from the rough endoplasmic reticulum (RER) into dual-membrane vesicles, in which the virus transcription and replication take place. The exoribonuclease (ExoN) function of nsp14 is a unique feature of coronaviruses, which supplies the correction capacity necessary to sustain a largescale RNA genome without accumulating harmful mutations. MERS-CoV and SARS-CoV transcribe 9 and 12 subgenomic RNAs respectively. These encode the four structural proteins, namely the spike protein (S), envelope (E), nucleocapsid (N),,membrane (M) and several accessory proteins that do not participate in viral replication but interfere with the host's innate immune response or whose function is unknown or misunderstood. The envelope “E” spike glycoprotein “S” clings to its cellular receptor, angiotensin converting enzyme 2 (ACE2) for SARS-CoV and dipeptidyl peptidase 4 (DPP4) for MERS-CoV. The “viral RNA genome” is delivered into the cytoplasm after membrane fusion, emissions to the host cell membrane or to the endosome membrane. The RNA is unwrapped to permit translation of the two polyproteins, transcription of the subgenomic RNAs and replication of the viral genome. The resulting envelope glycoproteins are introduced into the RER or Golgi membranes; genomic RNA and nucleocapsid proteins coming together to form the nucleocapsids. Virus particles bud in the ER-Golgi intermediate compartment (ERGIC). The virus-containing vesicles then fuse with the plasma membrane to deliver the virus [22]. Regarding COVID-19 (Fig. 3), genomic RNA is utilized as a scaffolding to directly translate polyprotein 1a/1ab (pp1a/pp1ab), which encodes non-structural proteins (nsps) to make the replication-transcription complex (RTC) in double membrane vesicles (DMV). Eventually, a nested set of subgenomic RNAs (sgRNAs) is synthesised by the RTC in a discontinuous mode of transcription. These subgenomic messenger RNAs (sgRNAs) have common 5′-leader and 3′-terminal sequences. and Subsequent acquisition and Transcription termination of a leader RNA occurs at transcriptional regulatory sequences situated between open reading frames (ORFs). These minus-stranded gRNAs act as a template for subgenomic mRNA production. A typical CoV genome and subgenomes include at minimum six ORFs. The first"ORFs (ORF1a/b), which account for about two-thirds of the total genome length, code for 16 nsps (nsp1-16), with the exception of Gammacorona virus, which has no nsp1. There is a − 1 frame shift between ORF1a and ORF1b, resulting in the production of two polypeptides: pp1a and pp1ab. These polypeptides are treated into 16 nsps by the virus-encoded chymotrypsin-like protease (3CLpro) or master protease (Mpro) and one or two papain-like proteases. Other ORFs on the third of the genome near the 3′-tip encode at least four major structural" proteins: (S), (M), (E) and (N) [33]. Various CoVs encode specific structural and accessory proteins,like the HE protein, the 3a/b protein and the 4a/b protein, in addition to these four basic structural proteins. CoV sgRNAs are used to translate all structural and accessory proteins [34].
Influenza A virus will turn hosts and create newly developed lineage [35]. This infection, known as zoonotic offers a chance to adapt the virus to the next host, and the resulting pandemics. When influenza A virus enters the body, the grippe HA (hemagglutinin) molecule accepts sialic acid (N-acetylneuraminic) around the top of the host cell [36]. HA is a transmembrane type of glycoprotein as homotrimer introduced to the virus sheet. Every monomer is composed of two subunits, HA2 and HA1. In the endosome low pH region, cleaved HA with a fusogenic HA2 stalk domain fusion mediates of the endosomal membrane with the viral membrane, which makes viral entry strong ribonucleoprotein (vRNP) to host cell [37]. The"vRNP complex comprises of 8 single-stranded, negative-sense nucleoprotein (NP) vRNA, and RNAs for influenza-A Polymerase"(PA, PB1, and PB2 compound)[28]. With subsequent fusion, the vRNP complex can be liberated in the cell's cytoplasm, after which it enters the nucleus by successful conveyance [37]. The nucleus is where the RNA synthesis of all influenza viruses takes place. Begin transcription process; RNA polymerase virus binds to highly retained and almost complementary 13 at the end of the 5' nuclear power and 12 at the end of the 3' nucleotide"eight segments. Nevertheless, the polymerase influenza virus has no inherent capping activity. It summarizes RNAs use 5 'host cap pre-mRNAs viral massager A special "cap snatching" mechanism mediation PB1 and PB2 derived from cellular transcriptions protein [28]. Nonstructural NS1 protein affects viral morphogenesis later in the viral replication cycle particles. However, they are not viral structural particles [38]. Receptor identification represents the initial stage of viral infection of host cells and one of the most essential factors in viral infection and pathogenesis [39]. While many other host- and virus-related factors can also influence the efficiency of infection and replication of the virus in a specific host, these factors only come into play once the virus is linked to a cell membrane receiver [40]. Coronaviruses (CoVs) have an enveloped, single-stranded-RNA genome that encodes four membrane proteins, namely spike, membrane, envelope, and nucleocapsid proteins as shown in Fig. 3 [2, 41]. S proteins are important for a viral entry concerning pathogenicity [42]. On the SARS-CoV envelope “E”, a trimeric S mediates the penetration of the virus into host cells. It first links to its host receiver, the angiotensin 2 converting enzyme (ACE2), and fuses the host and virus membranes afterward [43]. A given receptor-binding-domain (RBD) on the “S” SARS-CoV is sufficient to bind with high-affinity to ACE2 [44]. An important element in the pathogenesis of SARS-CoV and cross-species infections has been identified as the RBD/ACE2 binding affinity [45]. The experimental cross-reactivity of anti-SARS-CoV antibodies with 2019-nCoV spike proteins, that might have a significant consequences for the rapid manufacturing of antibodies and vaccines to combat 2019-nCoV, is therefore urgently needed [46].
Virulence factors are considered one of the vital element which plays a prominent role in virus adaption into the host cell [47]. Regarding pandemic influenza, Haemagglutinin (HA) is part of the surface glycoprotein of the virus. with two main roles in the very earliest phase of virus replication: membrane fusion and receptor linking [48]. HAs of high Avian influenza virus pathogen set an important contribution in virulence. They usually have a specific sequence (i.e. a sequence of basic amino acids at the cleavage site that helps to the prevalence of pathogenicity). [49]. This pattern, although it is not reflected in the HA sequence of 1918. Nevertheless, it had been shown that a reasserting virus with the genetic history 1918 HA repeated at a significantly elevated titer in the lungs, and with a large influx of lungs from neutrophils and alveoli macrophages caused severe pulmonary damage. The real 1918 virus with significant morbidity and subsequent death showed similar results. These results indicate an important role in the disease of the 1918 virus for the HA gene [50]. The high-virulence area (s) of HA has not yet been discovered. The other central factor in the 1918 outbreak is virulence. One unusual characteristic of the pandemic of 1918 was that many people have passed away of viral pneumonia; viral flu viruses in the pulmonary system of infected persons usually replicate poorly and often result in life-threatening viral pneumonia [51]. In the pulmonary system of infected filaments and non-human primates, we record an effective replication of the 1918 virus, which contributed to viral pneumonia. In comparison, the lungs of infected animals did not have a contemporary human H1N1 virus, even though it reproduced in the nasal cavities. Therefore, we conclude that the 1918 virus's capacity to expand in the lungs is related to its high human virulence [51]. It is also noted that high virus titers were based on the NP genes and polymerase in the lungs of infected ferrets in 1918 [52]. Polymerase genes also are significant in the pathogenicity and transmission of the mouse in ferrets [53]. These results strongly involve viral RNA polymerase complex in the successful transmission of the virus to the low respiratory tract and indicate that, in combination with a particular HA, it may be sufficient to induce fatal pneumonia during the pandemic of 1918–1919 [54]. Pathogenicity can also be correlated with other viral factors like the case of pandemic NA, NS1, PB1-F2 and others in 1918. The pro-apoptotic viral protein PB1-F2 needs only the shift in one amino acid at the 66th position to enhance the virulence of the virus in 1918 [51]. The 1918 PB1-F2 expression encourages pulmonary pathology in primary viral and secondary bacterial infections [51]. Respecting beta coronaviruses (SARS, MERS, and COVID-19), they contain E protein consisting of several active motifs between 76 and 109 CoV-dependent amino acids given its limited size [55]. Modification or suppression of E protein in different CoVs resulted in viruses with different phenotypes and unusual interrelationships between the virus and the host including stress induction and protein reactions or changes in concentrations with cellular ion because of E protein ion channel activity [55, 56]. All these practices have an important effect on the pathogenesis of CoV. Furthermore, in COVID-19 “S” is the main defining of cell tropism and therefore interspecies transmission of CoVs, since it binds the virus to a cellular receptor and then catalyzes membrane fusion entry of the virus [57]. The electron-microscopy 3D structure of the 2019-nCoV viral S showed its similarity with the S of the other COVs [57]. The further characteristics of other CoVs may thus be deducted. Viral S is a transmembrane type I transmembrane protein with an n-terminal cleavable signal peptide, a large and highly n-glycosylated e, a transmembrane, and cytoplasmic tail embedded in an S-cyllated residue cluster [57]. The ectodomain has been divided into the highly variable S1 domain between the genera and the S2 domain that is more conserved and catalyzes membrane fusion. The recipient-binding operation causes pathogenicity [58].
The pandemic H1N1 influenza virus in 1918–1919 witness variations in signs and symptoms according to many factors such as the severity of the cases, individual's age, and season [59]. The mild, unclear illness as predominate in the spring herald waves included symptoms of the upper respiratory tract, such as sore throat, nasopharyngitis, and cough, as well as systemic manifestations of fever, myalgia, and prostration (Fig. 4) [60]. Epistaxis has been carried out in both mild and severe cases [61]. The physician reporting Three thousand cases at Camp Fremont noted that epistaxis was a common characteristic of the entire pandemic [60]. This was considered a characteristic of the disease as blood always poured from the nose and mouth of the patient. The duration of moderate illness was generally limited to 72 h. Typically, the cough was not productive. Fever was prevalent up to 104° F. Sometimes sudden and extreme prostration [62]. One definition of the patient as "quickly or almost unexpectedly seized with a sense of prostration that was completely incapable of doing what he could. There was significant respiratory distress in patients with severe illness [62]. Their symptoms included remarkably intense cyanosis, hunger in the air, reduced awareness, and diffuse bubbling rales of highly progressive lung edema (Fig. 4). The cyanosis of heliotropic cyanosis found in some patients before death after the heliotrope flower's deep blue or purple color. Physicians also first note that the lips and ears are intensely blue before focusing on the rest of the face [31]. In a letter to a colleague's doctor, some described the color as purplish-black and one Scottish doctor who works at camp Deven noted that "the men color to white are not easy to distinguish." In the abnormal pigment, a doctor determined that cyanosis is due to extensive exudations in the alveoli preventing proper oxygenation when a repeated spectrographic control of the patient's blood was not found [60]. Two psychiatric conditions related, acute respiratory disturbance. Rapid mortality syndrome (ARDS), and fatal cases were identified of bronchopneumonia. A secondary bacterial infection leading to bronchopneumonia caused the most deaths with pneumonia, except for those killed in 1918 after the epidemic of H1N1 [60]. Initial leukopenia was followed by bronchopneumonia leukocytosis [63]. Brundage and Shanks [64], recorded a median period of 7–10 nights from the onset of illness and several deaths > 15 days after the onset, in conjunction with secondary bacterial pneumonia, for the most affected population. In the case of SARS, Fever, chilling, rigors, myalgia, dry toxins, dyspnea, malaise, and headache are the main distinguishing clinical features of SRAS [65]. More popular are sore throats, diarrhea, rhinorrhea, nausea, vomiting, and swelling [65]. In 40–70% of SARS patients, watery diarrhea was present. It tended to happen about 7 days after the onset of the disease [18]. Two patients complicated with epileptic status was detected in serum and cerebrospinal fluids. Elderly SARS-CoV infected patients may develop a low appetite, decrease in overall well-being, fall-fracture, and uncertainty, but may not be able to mount febrile responses to some elderly people [66]. In comparison, SARS-CoV infection was usually mild in kids under the age of 12. In contrast, infection in adult children was close to that in adults [67]. SARS-CoV-ac-Quired infection was connected to a lethality rate of 25%. during pregnancy, a high incident of spontaneous abortion, preterm delivery, and delayed development of the intrauterine child without perinatal SARS-CoV infection [18]. Adults who become infected with MERS-CoV may develop a range of illness and disease severity, from asymptomatic to slight, moderate or severe (Fig. 4) [68]. The time of incubation is from 2 to 14 days. Low-grade fever, runny nose, sore throat, dry cough and myalgia can occur in patients with mild infections. Patients with serious infections have acute pneumonia Syndrome of respiratory pain, a multilateral scheme organ failure, and disease. Furthermore, fellow members measure pneumonia progression by scoring the number of chest x-ray lung zones in patients with severe infection, showing abrupt progression after approximately seven days and severity pneumonia. The symptoms peaked afterward about fourteen days. MERS-CoV is greater in the lower respiratory tract samples than in the upper respiratory tract samples. Also, extrapulmonary characteristics are common, including myalgia. Back to half of all MERS-CoV patients, a third of critically ill people, including abdominal pains, nausea, vomiting, and diarrhea, are experiencing acute kidney injury. A third are gastrointestinal, and MERS-CoV in stool can be found [68]. Regarding COVID-19 fatigue and cough is myalgia or tiredness [69], most frequently reported symptoms. Expectoration, headache, haemoptysis and diarrhoea less frequent symptoms [70, 71], and in over half the patients, dyspnea developed (Fig. 4). The results of the blood tests showed that the white cell and lymphopenia were normal or reduced [72]. The typical ICU admission chest CT images were bilateral, multiple lobular, and sub-segment consolidation areas [73]. Non-ICU patients demonstrated bilateral ground-glass opacity and sub-segmental consolidation areas by representative chest CT conclusions [74]. Laboratory studies found that the most frequent symptoms are cough (67.7%), and fever (87.9%) while diarrhoea is uncommon.. 82.1% of ICU admitted patients reported lymphopenia [75].
The exact degree of pandemic morbidity and death from 1918 is not known because influenza does not differ from other respiratory diseases without laboratory confirmation [76]. Autopsy samples analyzed are mostly the pulmonary tissue of the fatalities that died in the autumn of 1918 [77]. There are also missing epidemiological details. The flu was not a reportable disease or illness monitored before the pandemic in any provincial or federal public health organization. After the pandemic death became apparent in the autumn of 1918, In addition, communities began to have physical offices to register cases of influenza [78]. Nevertheless, numerous cases avoided reliable reporting and/or timely reporting. In the American Public Health Journal of January 1919, the editor wrote in several cases that data was incomplete and deceptive because "the demand for intervention was so strong, that very few were willing to focus on research in the future [60]. Reported since the middle of the eighteenth century, significant pandemics have occurred between 10 and 40 years. Of these, the pandemic of "Spanish flu" in 1918 was the worst in the world, killing 20–40 million or more people (Fig. 5) [79]. In 2002 a SARS outbreak originating in Guangdong in China triggered the number-one major infectious disease of the twenty-first century, with 916 deaths among over 8098 patients in 29 countries. Ten years later, 2254 laboratory reported cases of MERS-CoV were declared by the WHO, with 800 deaths in 27 countries, from 2012 to 16 September 2018 [80] (Fig. 5). Significantly, over eighty percent of recent studies in virology and genetics of this infection have demonstrated that both MERS-CoV and SARS bats could be potential natural reservoirs. of confirmed cases of SARS, 22 percent were health workers in China and more than 40 percent were health workers in Canada [81]. Similarly, MERS nosocomial transmission was in the Middle East and in Korea. The cases reported in the Middle-East and North Africa have all contributed to the outbreaks in other countries, and their transmission due to international travel. Both SARS and MERS contributed to massive public health and economic outbreaks [80].
Dendritic cells (DCs) significantly contribute to innate immunity and can initiate large amounts of chemokines and cytokines [82]. These cells can move to the lymphoid tissue from peripheral tissue to activate the populations of T-cells [83]. On the other side, the key to immunity against viral infections is the adaptive T cells [84]; CD4 + T cells facilitate the virus-specific antibody production by T-dependent activation of B cells [85]. “CD8 + T cells” are nevertheless cytotoxic, killing viruses [86]. Inborn immune responses at the time of influenza can be defined as interactive between mucosal secretions and virions, epithelial-cell infections, and the activation of other types of residents of the epithelium or sub-epithelial layers [87]. Which involve lymphoid "innate" cells, resident macrophages, and dendritic cells (including alveolar macrophages) [87]. After the blood is recruited to the infection sites, other cell typologies like poly-morphonuclear leukocytes and monocytes take action [88]. Each cell type shows a different set of offers, they can sensitize the virus presence and enable special protective functions. Besides, the receptors are situated strategically in several subcellular components like cell membranes, endosomes, cytosol, and mitochondria [89]. This allows the host to mount defences tailored to the invading pathogen and take its tropism, intracellular lifestyle, and reproduction strategy into account [90]. For flu and other infections, therefore, innate responses may be conveniently divided into modules, each of which involves specific cell types, receptors, molecules of the effectors, and intracellular compartments. The components of the influenza module enclose (1) solvable extracellular proteins containing corporal fluids; (2) the interferon system; (3) different kinds of cytokines and chemokines able to orchestrate an innate response; (4) macrophage and neutrophils phagocytosis; (5) dendritic cell antigen presentation [91]. In the case of SARS infection, As the innate and acquired immune responses help the control of viruses and mild diseases, cytokine dysregulation, viral cytopathic symptoms, ACE 2 lung downregulation, irregular immune response and autoimmune processes lead to a more serious illness and eventual death, the progression of SARS may be linked to cell-mediated immunity from T-helper (Th1) and inflammatory hyper-innate response [92]. Significant enhancement in"Th1 and inflammatory cytokines (interferon-g [IFN-g], interleukin-1 [IL-1], IL-6, and IL-12), along with a significant increase in chemokines like Th1 chemokine, IFN-g, IL-10 inducible, (IP-10) neutrophil chemokine, and monocyte protein-1 chemokine attractions were observed over two weeks after the onset of the disease in a research study conducted in 20 adults infected with SARS-CoV [18]. The immune response mechanisms caused by MERS-CoV infection and immune evasion strategies have not been fully explored yet. Of particular interest, MERS-CoV evolved innate immunity control strategies and prevented or blocked the pathways of IFN production [93]. This skill can substantively be responsible for the high death rate levels of MERS-CoV patients, particularly those with immune-compromised. Once the virus has been recognized as Toll-like Receptors (TLRs), one of the two different adapter molecules is recruited either MyD88 (Myeloid Difference Primary Response 88) or Toll / Interleukin-1 Receptor-(TIR-) domain-containing Adapter-Inducing Interferon-β (TRIF). The molecules also activate the MAPK and NF-ŚB pathways which promote the development of pro-inflammatory retardants and IFNs"[93]. During COVID-19 infection both inborn and adaptive immune cells are synergistically involved in the anti-viral response. The rate of lymphocytes and subsets of T cells that act a significant role in regulating the immune response differs with potential viral pathological mechanisms depending on the type of virus. A considerable rise in the neutrophils, leukocytes and neutrophil–lymphocyte ratio (NLR) was seen in severe cases of COVID-19 compared to mild cases [96]. Prominent lymphopenia, which indicates an impairment of the immune system, occurs in most patients with COVI-19, especially in severe ones [97]. Therefore, it appears that leukocytes and neutrophils may strengthen the cytokine storm (CS) other than the COVID-19 lymphocytes. Past work has decreased the overall number of lymphocytes and T cells in patients with SARS-CoV infection [94]. The infection by SARS-CoV-2 may result in immune disorders dysregulation by affecting the T cell subsets [95–97]. Significant T cell alleviation is observed in COVID-19 and is more pronounced in severe cases. In COVID-19 patients, the levels’ cells (CD3+, CD8+), cytotoxic suppressor and helper T cells (CD4+, CD3+) and regulatory T cells are lower than normal levels. In contrast, helper T cells and regulatory T cells are remarkably lower in severe patients than in non-severe patients as shown in Fig. 6 [98]. Regulatory T cells are known to be responsible for maintaining immune homeostasis by suppressing the activation, proliferation, and pro-inflammatory function of most lymphocytes, including NK cells, T cells CD+4, T cells CD+8, and B cells (Fig. 6) [99]. Also, the percentage of naive helper T cells amplifies. In contrast, the percentage of memory helper T cells and CD28 + cytotoxic suppressor T cells decreases in severe COVID-19 cells [100]. The balance between the naive T cells and the memory T cells is fundamental to the effective immune response. Besides T cells, reduction of NK cells and B cells is observed in COVID-19 [101]. Overall, these results indicate that SARS-CoV-2 indicate that SARS-CoV-2 is responsible for immune misregulation with the induction of aberrant cytokine and chemokine responses [102], alteration of the lymphocyte subgroup, all of which could lead to cytokine storms and further tissue damage [103]. Excessive inflammatory response with a characteristic of cytokine storms cause serious illness and worsens the COVID-19 prognosis [104].
The age of an affected individual play a main role in identifying their risk of deceased during the 1918 influenza pandemic. In general, when seasonal influenza death rates are plotted as a function the age of the population, a "U" formed curve is obtained, with the highest level of mortality observed among the young and the elderly [6]. Conversely, pandemic epidemics (to varying degrees) are characterized by a shift in lethality towards younger age groups. This was particularly marked during the 1918 pandemic when young adults (15–30 years) generally had a high mortality rate that a "W" mortality curve was generated [6]. For SARS, Individuals of all ages were infected, and the median age was less than Forty-five years. Healthcare workers represented 22% of all cases in Hong Kong and 22.8% in Guangdong. In Canada and Singapore, the percentage of healthcare workers affected was higher, at 43% and 41%, respectively. The age and gender distribution of SARS in Hong Kong is as follows: 61.7% of patients are under 45 years of age, 21.2% are between 45 and 64 years of age, and the remainder are over 64 years of age. Eleven (8.14%) of the 135 early community cases with no history of contact with SARS patients were zoonotic type. The lethality rate in Hong Kong rises with age as in other world regions: 14.7% among people under 44, 21.4% between 45 and 64 and 63.9% over 64. Experience from Hong Kong and other areas suggests that deaths are linked to co-existing diseases in the oldest age group (> 64 years) [105]. In MERS, older adults and people with health conditions such as diabetes, chronic lung disease, kidney disease, or cancer. Additionally, patients with weakened immune systems, such as those receiving chemotherapy or immunosuppressant medication, and most of those who died of MERS had pre-existing chronic diseases [68]. According to the center for disease control and prevention (CDC), COVID-19 is a new disease with limited information on risk factors. Little information is available on the risk factors for serious illness. Based on clinical expertise and currently available pieces of information, adults and individuals of any age who have severe underlying medical conditions might be at higher risk for severe illness from COVID-19 [106]. So far, we know that people at high risk of severe disease from COVID-19 are 65 years and older and who lives in a nursing home or long-term care facility. Individuals of all ages with underlying medical conditions, mainly if not well controlled, including: People with severe to moderate chronic lung disease or asthma [107]. People who have serious cardiovascular problems [108]. People who are immune-compromised [109]. Numerous factors can make a person immunocompromised, including cancer treatment, smoking, bone marrow or organ transplants, immune deficiencies, poorly controlled HIV or AIDS, and prolonged use of corticosteroids and other immune-suppressing drugs. [110]. People with severe obesity (body mass index [BMI] of 40 or higher) [111]. People with diabetes [112]. People with the chronic renal disease under dialysis [113]. People with liver disease [114].
Pregnancy is a source of risk for disease and death [115]. This is associated with several physiological various transformations that take place during pregnancy. Due to the hormonal and mechanistic changes during pregnancy, several changes also happen in respiratory and cardiovascular systems, particularly increased stroke volume, heart rate, reduced O2 consumption, and decreased lung capacity (Fig. 7) [116]. Immunologically relevant changes also occur during pregnancy, shifting from cell-mediated to humoral immunity [117]. Such changes can make pregnant women more vulnerable or more likely to be seriously exposed to specific viral infections, such as influenza [118]. While suitable Control groups that are not pregnant are usually not as available, the death rates for pregnant women in the 1918 and 1957 pandemics appear to be abnormally high. Of the 1,350 influenza cases reported in pregnant women during the 1918 pandemic, the proportion of deaths was 27% [119]. Likewise, 45% of pregnant women hospitalized with influenza died in 1918, and 20% of pregnancy-related deaths during the 1957 pandemic; half of the deaths of women of reproductive age were due to pregnancy [120]. Additionally, in spring 1919, birth rates in all populations examined declined by an average of 2.2 births per 1,000 people, representing a 5–15% decline from baseline levels [120]. The low birth rate in 1919 reached its lowest point 6.1–6.8 months after the flu pandemic peak in the autumn, indicating that the birth shortfall was due to an excess of first trimester abortions among ten expectant mothers during the pandemic peak. Pandemic-related mortality was not sufficient to explain the observed results [121]. SARS coronavirus infection has been a contributing factor to severe maternal illness, maternal death and spontaneous miscarriage [122]. In a case–control investigation to determine the effects of SARS on pregnancy, ten pregnant and forty non-pregnant women with the infection were compared. In terms of symptoms, renal failure and disseminated intravascular coagulopathy were more common in pregnant women with SARS than in those without SARS. As a result, 60% of pregnant patients with SARS requested admission to an intensive care unit (ICU), compared with 18% of the non-pregnant group [122]. Another study reported by Maxwell et al. [123], based on a clinical study of a group of seven pregnant women infected with SARS-CoV, showed a mortality rate of 28%, compared with 10% for the non-pregnant positive control group. As a precautionary measure, SARS infected mothers have been recommended not to breast-feed their newborn children to prevent viral vertical transmission [122]. Most pregnant women displayed primary symptoms such as cough, dyspnoea and fever. The other main symptoms of COVID-19 in pregnant women are summarized in Fig. 7 [124–126].
The pandemic of COVID-19 has swept into over 200 countries with notable confirmed cases and deaths. It has caused mental health stress and public panic [127]. The rapid dissemination of COVID-19, which started in China, has been characterized as a pandemic by WHO in March 2020 [128]. In February 2020, Egypt announced its first COVID-19 case. After that, Egypt scaled-up precautionary measures, with a partial lockout beginning in March 2020 [129]. Reverse transcriptase-polymerase chain reaction (RT-PCR) was performed on nasopharyngeal swabs taken from symptomatic patients, and contacts of confirmed cases traced over the previous two weeks [130]. In individuals with a high suspicion rate, the test was repeated after 48 h [130]. The airport testing included body temperature and clinical assessment and the use of a rapid diagnostic test for severe-acute-respiratory syndrome IgM and IgG for coronavirus 2 [129]. The general public in China has severe anxious behaviors, resulting in a significant shortage of medical supplies. In addition, many frontline medical staff have been overworked for an extended period, which has prevented them from getting sufficient rest. There are indications that mental health problems may be occurring among health workers and survivors during the SARS epidemic [131]. However, the time distribution of the spread of COVID-19 may also commit to the heterogeneity of the disease burden across the US. The size and the time of the epidemiological peak, in particular, determine the required health system's responsiveness to provide proper healthcare. In many cases, it is challenging to obtain accurate data forecasts of the epidemic peak due to limited and often not trustworthy incidence data and the difficulties of modelling the effects of rapidly implemented and modified mitigation efforts [132]. Variability in county-level screening standards and actions, non-pharmaceutical interventions (NPIs) such as social distancing and outbreak initiation, and insufficient laboratory testing data also limit efforts to accurately model epidemic trajectories over several weeks [132]. In addition, projections of the cumulative burden of disease are less constrained by these difficulties. They do not attempt to describe the evolution of an epidemic over time. Even though these projections do not consider the nuance of the intensity and timing of outbreaks, their estimates of the disease burden's spatial footprint contain essential information for resource allocation [132].
During viral pandemics, various strategies have been used to limit the virus's propagation and treat infected and affected patients, such as quarantine, mass gatherings, facemasks, and hygiene [133].
For centuries, quarantine has been used to limit the emergence, introduction, transmission and spread of transmissible diseases [134]. When the “second wave” of Flu has been transformed into severe in 1918, Many nations have enforced stringent quarantine measures on all incoming carriers to prevent the propagation of the flu [6]. In most cases, these attempts failed. Restrictions have been put in place very late. The new viral infection was already well established in the country. The quarantine has been invoked by those infected who have not yet shown symptoms [135]. Thus, countries like the U K and S.A dismissed marine quarantine as inoperative and inefficient [6]. However, the Australian government imposed a naval quarantine before any victims of the 2nd wave were reported [136]. It helped protect Australia from the second pandemic wave until December 1918, when the quarantine was eventually finally over. The sea restrictions thus protected Australia from the pandemic. It indirectly helped protect some Pacific islands that were reliant on Australian supply boats. [137]. The most striking example is the difference in mortality between Western Samoa and American. A rigorous naval quarantine was enforced on American Samoa by the American Governor in 1918. This quarantine limited the flu from accessing the country, and no mortality cases from the 1918 flu were documented in American Samoa. This was in stark contrast to neighboring Western Samoa (∼ 100 km away), which did not have a strict sea quarantine. [6, 138]. Therefore, Western Samoa was infected by Tulane's New Zealand supply ship. Influenza is estimated to have claimed the lives of more than a quarter of the population [6]. In the case of SARS, prevention measures include early detection of cases and isolation, contact tracing and follow-up. Quarantine is efficient, but it is costly in time and resources and socially intrusive, so few countries can maintain such efforts over long periods [105]. Most quarantined individuals were confined to their homes during the SARS outbreak and active monitoring of symptoms [139]. In some countries, quarantine has been legally imposed and monitored by neighborhood support groups, the police, other workers, or home video cameras [139]. In other areas, consistency has been "requested", but judicial orders were handed down for a small percentage of non-compliant persons. According to reports, SARS was diagnosed in 0.2% of quarantined contacts in China-Taiwan, 3% in China-Hong Kong Special Administrative Region (SAR) and 4–6% in China-Beijing [139]. This is in part a result of different criteria for quarantining people. The most at-risk contacts (except for health care workers who have been exposed to certain unsafe conditions of patient care) was exposed to sick family members. Quarantine resulted in economic and psycho-social stress, risk communication, compensation and staffing issues for individuals, families, employers and governments [140]. Legitimate appeals and non-compliance with quarantine orders were rare [141]. Appropriate quarantine actions and measures that should be considered for MERS patients to prevent the spread of MERS-CoV to other susceptible individuals. Despite the fact that several studies in SK and SA have indicated that human-to-human transmission is relatively limited, it has been proven to be of very important, particularly in hospital outbreaks.[142, 143]. Accordingly, on 30 January 2020, Accordingly, WHO has determined that the coronavirus (COVID-19) disease is a public health emergency of international concern. As the epidemic continues to evolve, Member States are examining options to prevent the introduction of the disease into new areas or reduce human-to-human transmission in areas where the virus causing COVID-19 is already circulating [144]. Policy actions to reach these objectives may include quarantine, which implies restricting or separating from the rest of the population. These healthy people may have been exposed to the virus to monitor their symptoms and ensure early screening of cases. Many countries have the necessary legal authority to enforce quarantine [145]. The WHO advises that contacts of patients with COVID-19 confirmed in tests should be quarantined for Fourteen days after the last exposure to the patient. To apply quarantine, a contact with a person who is involved in any of the following activities from 48 h before until 14 days after the onset of symptoms in the patient (Fig. 8) [144]: Having direct contact with a COVID-19 patient from within one meter and for > fifteen minutes; Providing direct care to patients with COVID-19 without the use of appropriate personal protective equipment; Staying in the same nearby environment as a COVID-19 patient (including sharing a workplace, classroom or home or attending the same gathering) for a while; Travelling nearby (i.e. within one meter) to a COVID-19 patient, regardless of the means of transport used. Other situations, as indicated by local risk assessments. [144]. As a point of comparison, no quarantine has been issued in the United States for the recent SARS-CoV or MERS-CoV outbreaks. The last United States federal quarantine was imposed in 1963 to prevent or avoid the spread of smallpox from Sweden to the United States during a Swedish smallpox epidemic [3].
The pandemic of 1918 involved many cities implementing simple interventions to limit the propagation of the disease. In particular, restrictions were placed on many community gatherings where human-to-human transmission might occur [146]. As a consequence, schools, theatres, universities and dances have been closed. At the same time, mass gatherings, such as marriages and funerals, are banned to avoid overcrowding [6]. The peak mortality rate was lower in cities that rapidly implemented these non-pharmaceutical interventions within days of the first local cases than those that waited a few weeks to respond [147]. Overall mortality was also affected by the timing of the lifting of these interventions. For example, while restrictions on gathering people helped reduce transmission of the influenza virus, actual viral transmission resumed after those restrictions have been reduced (usually within 2–8 weeks of their implementation) [148]. Facial masks were a standard preventive measure used during the 1918 pandemic [149]. Although the infectious agent of pandemic influenza was not particular, the consensus was that it was an airborne disease and that wearing a face mask would prevent infection [150]. Consequently, many cities and regions, including Guatemala City, San Francisco and some prefectures in Japan, have made it mandatory to wear a medical proper face mask in public places, and special teams and education campaigns have been set up to enforce this regulation [6]. However, for a facemask to be at least marginally effective against the influenza virus, it must be: Used at all times, Correctly made and fitted, Composed of proper material [151]. SARS social distancing measures were implemented to protect from the fast-spreading of the virus, such as closing schools, cancelling mass gatherings, theatres, public facilities, public transport, restaurants or hospitals. Face masks in areas of suspected widespread unrelated community transmission of SARS coronavirus has been applied. [152]. Many people in these areas also opted to wear masks outside their homes. These measures were often implemented simultaneously with other actions, such as increased contact tracing, making their independent effectiveness challenging to assess. However, the simultaneous introduction of various criteria has been associated over time with a dramatic decrease in the number of new SARS cases [139]. A study in Beijing demonstrate that using a mask more frequently in public places may be associated with increased protection. Another case–control study in China and Hong Kong found that wearing a mask frequently in collective places, wash your hands more than ten times a day and thoroughly disinfecting living quarters appeared to offer protection [153]. Except for the group in Amoy Gardens, where accidentally produced sewage aerosols transmitted SARS-CoV, the transmission of SARS in the community from aerosols or in social settings appears to be uncommon [154]. Notwithstanding the multiple mass gatherings that provided millions of opportunities for the virus to spread, no outbreaks of MERS or MERS-CoV were reported during or immediately after these events. [155]. According to Hui et al., [156], some procedures used to enhance MERS control: Hand hygiene, droplet and contamination precautions for febrile patients before testing for MERS-CoV. Provide surgical masks to all patients undergoing haemodialysis and ensure that healthcare workers wear N95 filtering masks when caring for a patient with a confirmed MERS-CoV infection who is undergoing an aerosol-generating procedure. Patients suspected of being infected with MERS-CoV and admitted to dialysis or intensive care units should be placed in isolation rooms with a portable dialysis machine. Strengthen environmental cleaning and prevent non-essential staff and visitors from coming into contact with MERS-CoV infected patients. Regarding COVID-19, since March 2020, cancellations of international and national religious, sporting and musical events have increased as countries around the world take steps to avoid or minimize the spread of SARS-CoV-2 [157]. Numerous high-profile GMs have been cancelled or rescheduled, including sporting events like the Union of European Football Associations Euro 2020 football championship, the Formula One Grand Prix in China, the Six Nations rugby championship in Italy and Ireland, the Olympic boxing qualifiers, the Mobile World Congress in Barcelona, and the Umrah in KSA [157]. Actions to limit the propagation of COVID-19 must be prioritized based on their expected multiple on effective R divided by their cost. By this criterion, experimentation with and deployment of universal masks looks highly desirable. Facial masks help reduce community transmission when used with widespread screening, contact tracing, quarantine of potentially infected persons, hand washing, and physical distancing. Through their effect on R0, all of these measures have the ability to minimize the period of containment needed. As governments talk about relaxing lockdowns, maintaining transmissions at a low enough level will be essential to preserve health care capacity until a possible vaccine can be produced. Masks may be essential to preventing a “second wave of infections” from over-burdening the health care system. Further studies and scientific research are urgently needed to overcome this emergency. UNESCO declares that "where human activities may result in morally unacceptable harm that is scientifically plausible but uncertain, measures must be taken to avoid or mitigate such harm". This is known as the precautionary principle. The World Charter for Nature, adopted by the UN General Assembly in 1982, was the first international endorsement of the principle of prevention. It was implemented in an international agreement, called the Montreal Protocol of 1987. The Charter declares that the deaths and economic collapse that have already occurred as a result of COVID-19 are a morally unacceptable harm [158].
Despite spectacular advancements in medication drug therapy over the past decades, the causative agent of the 1918 influenza pandemic has been a mystery [159]. Without clear information on the agent responsible for the pandemic, a broad range of different therapeutic and preventive treatments have been inclines [160]. Individuals have experimented with drugs (including Aspirin) and home remedies such as mustard poultice, tobacco, beef tea, quinine, opium, saltwater, zinc sulphate inhalation, and alcohol [161]. As with the Japanese medicine Kampo (herbal remedies with green tea), the traditional Chinese medicine may provide a beneficial effect on the stimulation of perspiration (helping to lower fever), replacing lost fluids and improving vitamin C levels. Equally, when using traditional Chinese medicine may have reduced the illness's severity of the flu infections in some persons [6, 162]. Therefore, scientific studies are needed for the validation and determination of the active substances in order to be able to manufacture medicines based on these active substances on a large scale in the near future. Additionally, the dose-toxicity effect of these natural compounds needs to be seriously studied to prevent any potential side effects. Today, antiviral drugs are key factors in preventing and treating infection with the so-called influenza virus disease [163, 164]. In a regular influenza season, antiviral drugs are primarily which is used to cure or treat seriously ill patients, especially those with a weak immune system [164]. In the case of a pandemic, particularly in the period prior to a vaccination becoming available, anti-viral drugs are crucial to treat persons who have been infected and prevent infection among those who have been disclosed [165]. There are currently two drugs licensed for use against influenza, including adamantanes and NA inhibitors. Both rimantadine and amantadine are oral drugs that trigger the M2 ion channel of influenza A [166, 167]. The clinical use of these drugs is, unfortunately, is no longer recommended worldwide Because of the fact that the broad-based resistance of circulating influenza A viruses [167]. In particular, Inhibitors of NA target the enzymatic activity of the viral NA-protein [168]. Oseltamivir is administered as an oral oseltamivir phosphate, which is further transformed into its active carboxylate form in the liver [169]; Zanamivir is inhaled as a powder (which limits its use in people suffering from underlying respiratory problems) [170], and peramivir is administered intravenously, which is essential for persons who have been hospitalized [171]. The latter three drugs have been approved in the US, Australia, Europe, Canada, Japan, Korea, and Taiwan. They work by simulating sialic acid binding in the NA active site of A and B influenza viruses [169–171]. Both Zanamivir and Oseltamivir are efficient for prevention and post-exposure prophylaxis in persons [172]. However, these drugs have been randomized controlled trials in patients with less complicated influenza.; therefore, observational data should be used to monitor and evaluate the efficacy in critically ill and hospitalized patients, where the need is likely to be greater. Notwithstanding this limited scope, the results of the studies are systematically show improved outcomes from the use of NA inhibitors, including reductions in the incidence of pneumonia and hospitalization, and a reduction in the risk of hospitalized mortality. An additional consistency is that better results are obtained with early administration of NA inhibitors (within two days of the occurrence of symptoms). However, later administration may be of further benefit in critical situations [31]. The effectiveness of antiviral agents including ribavirin, INF, and protease inhibitors that was used to cure individuals with SARS-CoV infection in 2003 [173]. None of these therapies have proven benefit owing to a lack of prospective randomized, placebo-controlled clinical trial data. Supportive care continues to be the mainstay of treatment of SARS-CoV infection [173]. In the form of intravenous pulse methylprednisolone (MP), systemic corticosteroids were given to some patients with SARS-CoV infection for several reasons [174]. First, it was hypothesized that the clinical progression of pneumonia and respiratory failure associated with the peak viral load of SARS-CoV may be mediated by the host inflammatory response [175]. Treatment with systemic corticosteroid blocking agents considerably reduced MCP-1, IL-8, and IP-10 levels 5–8 days after treatment in twenty adults with SARS-CoV infection.[176]. In patients with fatal SARS-CoV disease, haemophagocytosis in the lungs was found, assigned to cytokine dysregulation. A therapy with systemic corticosteroids was therefore performed to modulate these immune responses [177]. However, prolonged use of systemic corticosteroid therapy may increase the risk of nosocomial infections, such as disseminated mycoses, metabolic derangements, psychosis, and osteonecrosis [178]. Recovering plasma, donated mainly by health care workers who had recovered entirely from SARS-CoV infection, appeared to be clinically helpful to the care other subjects with viral progression of SARS-CoV infection [179]. Delivery of convalescent plasma at an early stage appears to be more efficient and effective as, among eighty people who were infected with SARS-CoV received convalescent plasma at PWH, the discharge rate at day 22 was 58.3% for patients (n 5 48) treated within 14 days of onset of illness compared with 15.6% for those (n 5 32) treated beyond 14 days. In the absence of proven effective antiviral therapy, convalescent plasma and human monoclonal antibodies merit further investigation for the management of SARS-CoV if it returns [18]. Regarding MERS, various treatments already in existence and in development can be helpful antiviral such as ribavirin and mycophenolic acid (MPA) [180]. Currently, no specific treatments to treat ribavirin were empirically employed for serious of severe patients of MERS. However, there is no objective fact that they improve treatment performance [178]. Treatment with either lopinavir/ritonavir or IFN-b1b in the marmoset model was combined with better clinical, radiological and pathological results with lower viral loads compared to no treatment. In contrast, mycophenolic acid on its increases viral load and death rate [181]. In KSA, macrolide treatment usually begins before the patient arrives in intensive care [182]. In a study of 136 patients in a retrospective study, MERS patients, noted that macrolide treatment was not connected with reducing mortality or improvement in MERS-CoV RNA clearance [183]. A randomized controlled trial is underway in KSA. Comparative analysis of lopinavir/ritonavir, recombinant IFN-b1b, and standard supportive care against placebo and routine supportive care in patients with laboratory-confirmed MERS requiring hospital admission [184]. It was demonstrated that systemic corticosteroids delay viral clearance in critically ill patients with MERS-CoV infection.[185]. A range of anti–MERS-CoV drugs and host-directed therapies are considered potential therapies for MERS-CoV [183]. Antiviral and supportive treatments are clearly essential in treating patients with COVID-19 [186]. Because CS is frequently present in severe cases and is often the cause of the exacerbation, anti-inflammatory treatment can help prevent further aggravation [186]. As is well known, there are a number of types of anti-inflammatory medications, including non-steroidal anti-inflammatory drugs, chloroquine/hydroxychloroquine, immunosuppressant's, glucocorticoids, and inflammatory cytokines antagonists (such as TNF inhibitors, IL-6R monoclonal antibodies, IL-1 antagonists, Janus kinase inhibitors) [186, 187]. The use of corticosteroids may be justified in concert with the help of cytokine inhibitors such as anakinra (IL-1 receptor antagonist) or tocilizumab (IL-6 inhibitor). Intravenous immune globulin (IVIG) may also play a role in modulating an immune system that is in a hyper-inflammatory state [186]. Overall, the prognosis and recovery from this critical stage of illness are poor, and prompt recognition and application of such therapy may have the most significant yield [186, 188]. Table 1 displayed different approaches for the treatments against COVID-19 with the reaction mechanism.
Vaccinating is one of the world's most successful ways to prevent disease, as indicated by the WHO. [192].”A vaccine helps the body’s immune system recognize and fight pathogens like viruses or bacteria, which then keeps recognize and fight pathogens like viruses or bacteria, which then recognize and fights pathogens such as viruses or bacteria, which protect us from the diseases they cause “[193]. Vaccinations protect from more than twenty five debilitating or life-threatening diseases, including polio, measles, tetanus, diphtheria meningitis, flu, typhoid and cervical cancer [194]. Nowadays, most children receive their immunizations on time. However, nearly twenty million people worldwide still miss outputting them at risk of serious diseases, death, disability, and ill-health [195]. First inactivated influenza vaccine was mono-valent (influenza A) [196]. In 1942, a bi-valent vaccine was produced after discovering the influenza B virus. It was later identified that the influenza viruses mutate, leading to antigenic changes [197]. WHO has published annual recommendations since 1973 for the influenza vaccine composition based on the results of systems of surveillance that identify currently circulating strains [198]. In 1978, the first trivalent vaccine included two strains of influenza A and one strain of influenza B. Currently, two strains of influenza B are circulating; the most recent WHO guidance propose adding a second B strain to make a quadrivalent vaccine [197]. Moreover, currently available inactivated seasonal influenza vaccines may even prevent the induction of cross-reactive CD8 + T-cell responses, which are our primary protection in a pandemic. They may therefore prove to be a double-edged sword [199]. Prompt production of vaccines also remains a challenge for future influenza pandemics. This was particularly evident during the 2009 pandemic, when sufficient quantities of pandemic vaccine were not available until October 2009, well after the pandemic had spread worldwide [200]. Vaccine production can be even more complicated because certain avian influenza viruses can cause the death of embryonated chicken eggs needed for vaccine production..[201]. Different vaccine strategies are required to accelerate vaccine production and overcome these problems. Nevertheless, An influenza vaccine that provides broad-spectrum, long-lasting immunity remains the gold standard for pandemic planning [6]. Continued research is needed to understand how a universal influenza vaccine can be implemented. In SARS, S protein ensures an essential function in the regulation of the viral infection through binding receptors and membrane fusion between the virus and the target cell [202]. An adenovirus-vaccine-based can stimulate potent SARS-CoV-specific immune responses in rhesus macaques and is promising for the establishment of a vaccine to combat SARS-CoV [203]. Other researchers have pointed out that the gene S DNA vaccine can induce the expression of specific IgG antibodies to SARS-CoV effectively in mice, with a seroconversion rate of 75%.after three immunization doses. In contrast, virus replication was decreased by over six orders of magnitude in the respiratory tracts of mice injected with S-plasmid DNA expression vectors. Protection was provided by a so-called humoral immune mechanism [204, 205]. The recombinant S protein showed antigenicity and receptor binding capacity. In contrast, synthetic peptides that elicit specific antibodies to the S-CoV S protein could be an alternative approach to SARS vaccine development [202, 206]. There is currently no vaccine that can protect against MERS-CoV infection. Many research groups are working on developing a using various platforms and several strategies, and some have shown their effectiveness in animal models [183]. Vaccination is perhaps the preferred choice for controlling COVID-19 [207, 208]. Epitopes, mRNA, and S protein-RBD structure-based vaccines have been widely proposed and started [209]. Rapid reconstruction of SARS-CoV-2 using a synthetic genomics platform has been reported, and this technical advance is helpful for vaccine development. [210]. The human ACE2 and rhesus monkey transgenic mouse models of COVID-19 have been well established for vaccine development [211]. A number of SARS-CoV-2 vaccines are already in ongoing clinical trials [205].
Tuberculosis vaccine Bacillus Calmette-Guérin (BCG) is a lively attenuated vaccine developed at the beginning of the twentieth century at the Pasteur Institute in Paris [212, 213]. Since that time, it has been the most widely used vaccine. Globally, with approximately one hundred and thirty million children being vaccinated each year [214]. However, it is interesting to note that shortly after its first introduction in Europe in the nineteen-twenties, epidemiological studies indicated that BCG vaccination greatly reduced infant death rate [212]. More recently, BCG vaccination has been shown to be correlated with reduced case death rate for COVID-19. The latest data from publicly available resources also indicate that the incidence of COVID-19 and the total number of deaths are strongly associated with the presence or absence of national mandatory BCG vaccination programmers [215]. On the basis of clinical results and experimental data, it is assumed that BCG induces long-lasting immune system changes that lead to enhanced responses to infections in both innate and adaptive immunity. [216]. In innate immune cells, BCG induced histone modifications and epigenetic reprogramming at the promoter sites of genes coding for inflammatory cytokines such as interleukin (IL)-1, IL-6 and tumor necrosis factor (TNF). This process has been termed "trained immunity" [217]. "In two studies, BCG was evaluated in Japan and BCG in Denmark for inducing cytokine secretion in peripheral blood lymphocytes"[218]."One study, carried out in Africa, demonstrated that BCG Japan caused more robust proliferation of CD4 + and CD8 + T cells, higher secretion of Th1 (interferon-c, TNF-a and IL-2) and lower secretion of Th2 cytokines (IL-4) compared to BCG Denmark [218]. Another study in Mexico showed that "BCG Japan induced higher levels of IL-1a, IL-1b, IL-24 and IL-6 in peripheral blood mononuclear cells obtained from vaccinated children, compared to BCG Denmark"[219]."These results suggest that BCG Japan is more effective than BCG Denmark in inducing the production of several types of inflammatory cytokines"[215].
Intravenous immunoglobulin (IVIg) therapy decreases intestinal epithelial cell infectivity. It reduces the growth of the opportunistic Candida albicans (human unicellular-fungal pathogen) in the murine gut in connection with the upregulation of anti-inflammatory cytokines coupled with downregulation of proinflammatory mediators [220]. CoVs cover positive-stranded-RNA viruses relating to the Coronaviridae family [221]. Viral RNA genome sequencing has shown that the virus-producing COVID-19 is phylogenetically linked to the SARS-associated CoV initial separated in Chinese horseshoe bats through 2015–2017. CoVs are extremely deadly with human-to-human communication, which has grievously-produced several losses. Unluckily, we were collectively disappointed to know that the opportunity was apparent, imminent, and important. We declined to ensure that the policies and directives, produced by specialists, can be quickly and efficiently performed in case of a disorder [222, 223].
Intravenously introduced nanomaterial of natural and inorganic sources is growing attention in clinical sciences [224]. Macrophages (a kind of immune cell covering cellular waste, bacteria, and other unknown particulate materials) of the spleen and the liver immediately prevent blood-borne bits. This is doubtful if the preferred target for healing NMs is the preferred target remains outside human organs [225]. Still, the adsorbed and functionalized blood protein on the outsides of NMs, but not the polymer layer, accidentally and predominantly induced activation of the other pathways [226]. On cover adsorption, plasma proteins may support conformational and thermodynamic differences and consequently, display reactive combinations that may convert responsive to C3b initiative. Ultimately, this makes C3bBb and C3bBbP convertases connected to cover-adsorbed protein. This method may be explained as a non-specific general mechanism by which NMs, despite their chemical structure and composition, could trigger the other pathway of the whole way. Moreover, it is stated that protein-C3b and protein-C3 convertases are produced and delivered from the NMs outside. [227]. As C3b is an opsonic fragment, connected C3b coupling and relief (in the frame of protein-C3b) may reveal why long-circulating NPs are gradually accepted and removed from the human blood by the macrophages of the spleen and liver.
The antimicrobial characteristics of some NPs like Ag NPs are quite recognized [228]. The importance of various NPs is being FDA approved for complete wound closure. It is possible for the antimicrobial colloidal silver compounds, passed by breath, to reduce the inflammation of respiratory system diseases. It is proven that Ag NPs have antimicrobial and antiviral characteristics [229]. However, there is no accurate study about the possibilities of breath application of NPs for the restriction and/or processing of respiratory diseases. Therefore, before the safe use of Ag NPs, several steps should be conducted (a) defining optimal Ag NPs metal features for most maximum powerful anti-viral structure, (b) expected useful inhibitory concentration to be taken at the objective respiratory system area, and (c) the necessary dosage for effective implementation by breath control. To determine the necessary delivery dosage, an advanced system must be designed to assess the impact of all platforms among the aerosol stocks by the aerosolizing material to the ultimate tissue displacement. In the context of breaths, the Ag NPs solution is proper. The critical opening point for determining any adequate dosage is to set the necessary point inhibitory concentration of the active factor. The standard anti-viral potential is achieved with NPs with less than 10 nm. In production, the NP's size stabilization (stabilizing or capping agents) process changes the anti-viral power. The current gum acacia capping mainly hinders the antiviral impact. Furthermore, it seems that Ag NPs of size less than 10 nm have more powerful anti-viral potential than Ag NPs with size more than 25 nm [230]. Communication among different NPs and the immune system becomes significant, and there are foundational issues regarding the security of the synthesized NPs. NPs can interact with many vital elements (cells, proteins, receptors, etc.) or activate cell signaling pathways, and subsequently, create variable immune replies (suppression and/or activation) and also severe medical conditions (cancer and/or autoimmune diseases) [231]. Directed NPs can be created to individually-associate with or withdraw identification by the immune system. Synthetic NPs have been used regularly to make new immunotherapy approaches. Immunotherapy includes the intentional infliction of the immune system as a therapeutic approach [232]. One of the main forces of immunotherapy is that there can be limited harmful side outcomes than those connected with conventional treatments [233].
The surface of NPs can be altered with special receptors to be attached with particular targets. In the innate immune system, phagocytosis relies on the stability of the signals of prophagocytic and antiphagocytic on target. For example, the copper oxide (CuO) NPs approach followed in the up-regulation of heat-connected proteins and stimulated ROS. Also, gold nanoparticles (Au NPs) caused up-regulation of the provocative mediator (NF-κB) that is negotiated within dysregulation of the immune homeostasis after stopping the function of the TIPE2 (tumor necrosis factor-α-induced protein 8-like 2) protein [234]. Proinflammatory cytokines may be caused by Toll-like receptors (TLR) indicating pathways. Several tumor necrosis factor and cytokines can stimulate inflammatory groups, improve the vascular permeability, and produce inflammation during severe inflammatory responses [235]. Cytokines are primary mediators of heat release (fever) [236]. TNF-α stimulates endothelial cells starting to hypotension. Some directed NPs can also stimulate inflammasome signaling pathways [237, 238]. Titanium dioxide (TiO2) NPs and crystalline silica (SiO2) NPs produce the inflammasome and cytokine discharge in bone marrow-acquired macrophages [237, 238] as displayed in Fig. 9. Peeters et al. lately stated that SiO2 stimulated inflammasomes in the lung epithelial cells and basic bronchial epithelial cells, which increased the inflammatory signal and induced fibroblast generation. Ag NPs effected inflammasome development and triggered cytokine discharge [239]. Inflammasome-stimulation-related cytokine creation by dendritic cells in reply to particle treatment was size-subject. The activity was achieved when the synthesized NPs were between 500 and 900 nm. Yazdi et al. stated that SiO2 NPs and TiO2 NPs, stimulate the inflammasome, leading to cytokine discharge [240]. The NP's specificity for selecting and invading tissues in symptomatic imaging and drug-based treatments is essential to block non-specific cell junction in normal human tissue. The organization of the nano-formulations has tried to overcome this force by intense, magnetic targeting. Some NPs have been utilized for diagnostics, several usually Au NPs, and magnetic NMs. Au NPs for diagnostics DNA small sections may be connected to Au NPs with a 13.0 nm small diameter NPs. These connections over a sensor outside occur after a corresponding target. Au NPs are particularly-efficient designs for sensors due to various scientific methods that may be utilized to identify them [241]. Table 2 summarized nanomaterials that have been used for COVID-19 treatment, diagnosis and immune system boosting with their reaction mechanism.
Nanotoxicity is an evolving field used to assess the unintended hazardous effects of nanoparticles (NPs) on human health. While NPs have become promising tools for a wide range of biomedical applications, their extensive use depends on the assessment of their biosafety [245, 248, 258–261]. There is increasing interest in assessing the health impact of these materials and expanding knowledge of their cytotoxicity and biocompatibility. Once a new nanomaterial appears, its cytotoxic effect, i.e., the possible alteration of basic cellular functions, is usually evaluated primarily. Nevertheless, the absence of cytotoxicity does not confer to these materials an implicit biocompatibility [262]. This must be assessed as a separate endpoint. The concept of biocompatibility is based on the adequate interaction between the nanomaterial and its biological environment, i.e. the non-existent toxic or immune response of the treated biomaterial (cell, tissue or organism) [262]. Cytotoxicity is generally related to the possible negative impact on a specific cell line. Thus, cytotoxicity is generally assessed first by specific tests conducted in vitro before being assessed in vivo. These tests can be performed in vitro (metabolic activity, cell proliferation and viability, oxidative stress, apoptosis tests, necrosis tests, etc.) and in vivo (behavioral analysis and body weight, biodistribution, biodegradation and clearance, pharmacokinetics, hematology and serum chemistry, histopathology, acute and repeated dose toxicity, reproductive and developmental toxicity, genotoxicity and mutagenicity, etc.) tests [262]. Regardless of whether in vitro or in vivo methods are used, the results of studies conducted on the toxicity of NPs are currently contradictory. Basically, it has been observed that cytotoxicity and biocompatibility are governed by several factors, including the inherent physicochemical properties of nanoparticles and the way they are delivered to the body. For example, a higher toxicity of nanoparticles was observed during oral and intraperitoneal administration compared to intravenous injection. In addition, biocompatibility was highly tissue or organ specific. The cytotoxicity of nanoparticles is also strongly related to physicochemical characteristics such as size, shape, surface area, and charge. All these elements show that the biocompatibility of nanoparticles is highly dependent on several factors, ranging from the intrinsic properties of the particles to the formulation, the biological target, the dose and even the methodology used to assess their toxicity [263]. In general, the smaller the size of the nanoparticles, the greater the cytotoxic effect. One of the hypotheses that could explain the toxicity of small-sized nanoparticles is that it results from the presence of a high surface area compared to their volume. This leads to an increased absorption capacity and may increase the risks of interaction with biomolecules [262, 263]. In addition, the dose of NPs administered to a model organism can also affect their toxicity. Similarly, Coradeghini et al. [264] revealed that the toxicity of Au NPs towards mouse fibroblast cell lines was dose-dependent. Recently, Donskyi et al.[265], proposed graphene with precise double sulfate/alkyl functionalities as a platform for the inhibition of SARS-CoV-2 and feline coronavirus replication by virtue of viral envelope disruption. Notwithstanding using a wide concentration window (10 to 100 times), the graphene platforms show strong antiviral activity against native SARS-CoV-2 without significant toxicity against human cells. Hence, more research is needed on the effective doses and likely toxic effects of NPs to create a safe environment for humans against extremely dangerous diseases like COVID-19. Therefore, new specific standardization and certification tests (which include evaluation of physicochemical characteristics, sterility, pyrogenicity, bio-distribution and ADME, pharmacokinetics, and in vivo and in vitro toxicity) for preclinical nanosafety and toxicity risk study need to be developed. Efforts to standardize risk assessment procedures for NPs are ongoing and need to be further improved. Currently, nanomaterials are considered in the same way as conventional chemicals. The main efforts to standardize nanotechnology are being developed by the standards development organizations (SDOs).
Biodegradable nanoparticles (BNPs) are novel carriers for the delivery of drug molecules. They are receiving increasing attention due to their ability to serve as viable carriers for the specific delivery of vaccines, genes, drugs and other biomolecules to the body (Fig. 10). BNPs have become popular recently due to their special features such as targeted drug delivery, better bioavailability and therapeutic efficacy to deliver the drug at a constant rate. They offer improved biocompatibility, superior drug/vaccine encapsulation and convenient release profiles for a number of drugs, vaccines and biomolecules for use in a variety of applications in the field of nanomedicine [266–269]. Polymeric nanoparticles are considered as biodegradable material. They are polymeric colloidal elements of very small size in which a drug of interest can be encapsulated or incorporated into their polymeric network or conjugated or adsorbed onto the layer. Various natural and synthetic polymers are used in the synthesis of biodegradable nanoparticles (Fig. 10), some of the frequently used polymers are chitosan, cellulose, gelatin, gliadin, polylactic acid and polylactic-co-glycolic acid. Nanoparticles have been progressively explored for drug delivery and have enabled sustained kinetic release. Drugs integrated in this system can give better efficacy, decrease drug resistance, reduce systemic toxicity and symptoms, and also improve patient compliance [268]. Recently, Qiao et al. [270], proposed a peptide-based subunit candidate vaccine against SARS-CoV-2 delivered by biodegradable mesoporous silica nanoparticles induced high humoral and cellular immunity in mice. Through this study, seven linear B-cell epitopes and three CD8 + T-cell epitopes were selected from the SARS-CoV-2 spike glycoprotein by immune computational approaches for vaccine design. A nanoparticle-based candidate vaccine (B/T@BMSNs) against SARS-CoV-2 was promptly prepared by encapsulating these ten epitope peptides into BMSNs, accordingly. The BMSNs, endowed with potential biodegradability and excellent in vitro and in vivo safety, demonstrated the ability to efficiently deliver the epitope peptides into the cytoplasm of RAW264.7 cells. Strong humoral and cellular Th1-like immunity was induced by B/T@BMSNs in mice and the 10 selected epitopes were identified as effective antigenic epitopes capable of inducing a robust peptide-specific immune response [270]. To optimize NPs as a delivery system, a better understanding of the different mechanisms of biological interactions and particle engineering is still needed. However, biodegradable NPs seem to be a promising system for drug delivery due to their versatile formulation, sustained release properties, subcellular size, and biocompatibility with various tissues and cells of the body. The development of nanomedicine remains a great challenge and other biodegradable nanomaterials still need to be explored and validated for their potential clinical use. Figure 10 provides a summary of the synthetic routes for biodegradable nanoparticles and their in vivo applications.
During the years 1919, 2002, 2012 and 2019, the world was attacked by four viral respiratory diseases, Spanish flu, SARS, MERS and COVID-19, respectively. Coronaviruses are single-stranded, non-segmented, enveloped, RNA-positive viruses that have a particular appearance under negative-stain electron microscopy. It is well known that wild waterfowl are the source of all influenza viruses in other species. Regarding COVID-19 fatigue and cough is myalgia or tiredness, most frequently reported symptoms. Sputum, headache, hemoptysis, and diarrhea were less frequent symptoms, and in over half the patients, dyspnea developed. During COVID-19 infection both inborn and adaptive immune cells are synergistically-involved in the anti-viral response. A significant increase in neutrophils, leukocytes, and neutrophil lymphocyte ratio has been identified in serious or critical cases of COVID-19 compared to mild cases. From currently available information and clinical expertise, the elderly and people of all ages with serious underlying health problems may be at increased risk of severe illness from COVID-19. The occurrence of pregnancy has been a strong risk factor for increased illness and mortality for both pandemic and seasonal influenza. During viral pandemics, a variety of different approaches were employed to limit or avoid the fast spread of the virus and treat infected patients such as quarantine, mass gatherings, facemasks, and hygiene. Today, anti-viral drugs are key in preventing and treating influenza virus infection and disease. Different medicines and approaches had been used for the treatments against COVID-19 such as remdesivir, favipiravir, ribavirin, chloroquine and hydroxychloroquine, glucocorticoids, teicoplanin and other glycol-peptides, monoclonal or polyclonal antibodies, convalescent plasma, and herbal medications. Vaccination is one of the most powerful ways for disease control and prevention, according to the WHO. A vaccine has the effect of helping the body's immune system to recognise and fight pathogens including the following viruses or bacteria, which then keepsrecognize and combat pathogens such as viruses or bacteria, which then recognize and fights pathogens like viruses or bacteria, thus protecting us from the diseases they induce. Intravenously introduced nanomaterial of natural and inorganic sources is growing attention in clinical sciences. The standard antiviral potential is achieved with NPs with less than 10 nm. In production, the NP's size stabilization (stabilizing or capping agents) process changes the anti-viral power. The NPs surface can be altered with special receptors to be attached with particular targets. In the innate immune system, phagocytosis guided by the stability of the signals of prophagocytic and antiphagocytic on target. This review presents insights about using NMs to give treatment to COVID-19 further, improve the bioavailability of the abused drugs, diminish their toxicity, and improve their performance. | true | true | true |
|
PMC9590114 | Li Song,Qiankun Li,Yao Lu,Xianqi Feng,Rungong Yang,Shouguo Wang | Cancer Progression Mediated by CAFs Relating to HCC and Identification of Genetic Characteristics Influencing Prognosis | 15-10-2022 | Background Hepatocellular carcinoma (HCC) is one of the most common malignancies, and although there are several treatment options, the overall results are not satisfactory. Cancer-associated fibroblasts (CAFs) can promote cancer progression through various mechanisms. Methods HCC-associated mRNA data were sourced from The Cancer Genome Atlas database (TCGA) and International Cancer Genome Consortium (ICGC) database. First, the differentially expressed CAF-related genes (CAF-DEGs) were acquired by difference analysis and weighted gene coexpression network analysis (WGCNA). Moreover, a CAF-related risk model was built by Cox analysis. Kaplan-Meier (K-M) curves and receiver operating characteristic (ROC) curves were utilized to evaluate the validity of this risk model. Furthermore, enrichment analysis of differentially expressed genes (DEGs) between the high- and low-risk groups was executed to explore the functions relevant to the risk model. Furthermore, this study compared the differences in immune infiltration, immunotherapy, and drug sensitivity between the high- and low-risk groups. Finally, we verified the mRNA expression levels of selected prognostic genes by quantitative real-time polymerase chain reaction (qRT-PCR). Results 107 CAF-DEGs were identified in the HCC samples, and five prognosis-related genes (ACTA2, IGJ, CTHRC1, CXCL12, and LAMB1) were obtained by Cox analysis and utilized to build a CAF-related risk model. K-M analysis illustrated a low survival in the high-risk group, and ROC curves revealed that the risk model could accurately predict the 1-, 3-, and 5-year overall survival (OS) of HCC patients. In addition, Cox analysis demonstrated that the risk score was an independent prognostic factor. Enrichment analysis illustrated that DEGs between the high- and low-risk groups were related to immune response, amino acid metabolism, and fatty acid metabolism. Furthermore, risk scores were correlated with the tumor microenvironment, CAF scores, and TIDE scores, and CAF-related marker genes were positively correlated with all five model genes. Notably, the risk model was relevant to the sensitivity of chemotherapy drugs. Finally, the results of qRT-PCR demonstrated that the expression levels of 5 model genes were in accordance with the analysis. Conclusion A CAF-related risk model based on ACTA2, IGJ, CTHRC1, CXCL12, and LAMB1 was built and could be utilized to predict the prognosis and treatment of HCC. | Cancer Progression Mediated by CAFs Relating to HCC and Identification of Genetic Characteristics Influencing Prognosis
Hepatocellular carcinoma (HCC) is one of the most common malignancies, and although there are several treatment options, the overall results are not satisfactory. Cancer-associated fibroblasts (CAFs) can promote cancer progression through various mechanisms.
HCC-associated mRNA data were sourced from The Cancer Genome Atlas database (TCGA) and International Cancer Genome Consortium (ICGC) database. First, the differentially expressed CAF-related genes (CAF-DEGs) were acquired by difference analysis and weighted gene coexpression network analysis (WGCNA). Moreover, a CAF-related risk model was built by Cox analysis. Kaplan-Meier (K-M) curves and receiver operating characteristic (ROC) curves were utilized to evaluate the validity of this risk model. Furthermore, enrichment analysis of differentially expressed genes (DEGs) between the high- and low-risk groups was executed to explore the functions relevant to the risk model. Furthermore, this study compared the differences in immune infiltration, immunotherapy, and drug sensitivity between the high- and low-risk groups. Finally, we verified the mRNA expression levels of selected prognostic genes by quantitative real-time polymerase chain reaction (qRT-PCR).
107 CAF-DEGs were identified in the HCC samples, and five prognosis-related genes (ACTA2, IGJ, CTHRC1, CXCL12, and LAMB1) were obtained by Cox analysis and utilized to build a CAF-related risk model. K-M analysis illustrated a low survival in the high-risk group, and ROC curves revealed that the risk model could accurately predict the 1-, 3-, and 5-year overall survival (OS) of HCC patients. In addition, Cox analysis demonstrated that the risk score was an independent prognostic factor. Enrichment analysis illustrated that DEGs between the high- and low-risk groups were related to immune response, amino acid metabolism, and fatty acid metabolism. Furthermore, risk scores were correlated with the tumor microenvironment, CAF scores, and TIDE scores, and CAF-related marker genes were positively correlated with all five model genes. Notably, the risk model was relevant to the sensitivity of chemotherapy drugs. Finally, the results of qRT-PCR demonstrated that the expression levels of 5 model genes were in accordance with the analysis.
A CAF-related risk model based on ACTA2, IGJ, CTHRC1, CXCL12, and LAMB1 was built and could be utilized to predict the prognosis and treatment of HCC.
Liver cancer is one of the commonest malignancies. In accordance with the Global Cancer Statistics 2020, liver cancer is the 6th for incidence and 3rd in mortality among malignancy-related deaths [1–3]. Secondary, liver cancer includes hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC), of which HCC accounts for about 75-85%. Although various options such as chemotherapy with sorafenib, surgical resection, and liver transplantation are applied in treating HCC, but there is still a poor overall prognosis, with an overall survival (OS) of 3-5% [4–6]. Therefore, it is essential to find available targets for HCC treatment [7]. Cancer-associated fibroblasts (CAFs) can secrete growth factors, cytokines, and inflammatory ligands, which stimulate epithelial-mesenchymal transformation (EMT), promote tumor proliferation and migration, and induce therapy resistance and immune exclusion [8–10]. Studies showed that CAFs engaged in bidirectional signaling with liver progenitor cells and can act as cancer stem cells, suggesting a close link between cirrhosis and liver cancer development [11]. In addition, CAFs support tumor growth in the liver. For example, CAFs can influence tumorigenesis by altering ECM stiffness. For example, CAFs can influence tumorigenesis by altering ECM stiffness; moreover, the cytokines and other factors secreted by CAFs may promote tumor growth, tumor angiogenesis, and epithelial to mesenchymal transition (EMT) [12]. In this study, samples in the TCGA dataset were grouped into high CAF/low CAF score groups with CAF scores, and then, 107 differentially expressed CAF-associated genes (CAF-DEGs) were utilized for risk regression analysis. Furthermore, 5 prognostic genes were gotten and utilized to establish a risk model, which provided a reference for applying CAF-associated genes (CAFGs) in the clinical prognosis and treatment outcome of HCC.
The mRNA expression data of 50 normal and 371 HCC samples, of which 360 HCC samples have available survival data, were sourced from The Cancer Genome Atlas database (TCGA). The mRNA expression data of 243 HCC samples were acquired from the International Cancer Genome Consortium (ICGC) database as a validation set.
xCell can calculate the abundance of various cells based on the single-sample gene set enrichment analysis (ssGSEA), which includes cancer-associated fibroblasts [13]. This study counted the mass of 21 immune cells in 421 samples of TCGA-HCC dataset by xCell. The samples were grouped into high and low CAF with the median number of CAF cells. Kaplan-Meier (K-M) survival analysis was performed based on the high and low CAF groups and the survival information of the HCC samples. Then, we collated the clinical traits of the samples, STAGE subgroups, and GRADE subgroups and compared the differences in the proportion of CAF cells between the STAGE subgroups and GRADE subgroups using chi-square tests.
The genes with similar expression patterns can be gathered, and the module that was highly correlated with traits can be filtered by WGCNA, thus finding the target genes relevant to the study [14]. To further identify CAFGs, we performed a WGCNA analysis. First, we clustered the 371 HCC samples to see the overall correlation of all samples in the dataset. The soft threshold was determined to ensure that the interaction between genes maximally conformed to the scale-free distribution, and then, the coefficient of dissimilarity between genes was introduced based on the adjacency between genes, and the systematic clustering tree between genes was obtained accordingly. Similar modules analyzed by the dynamic tree cutting algorithm were merged (MEDissThres = 0.2). Finally, the correlations between the modules and CAF were calculated, and the key modules were selected with the criteria of |cor| > 0.4, p < 0.05. Moreover, the genes in the key modules were the CAFGs.
We performed a differential analysis in the TCGA dataset for high CAF samples and low CAF samples to obtain differentially expressed genes (DEGs) between high and low CAF samples and differential analysis for normal and HCC samples. The screening condition for the differential analysis was p adjust. < 0.05 and |log2FC| > 0.5. To identify CAF-DEGs, we crossed CAFGs, DEGs between high and low CAF, and DEGs between normal and HCC samples.
In this study, 360 samples containing survival information in the TCGA dataset were grouped into a training set and a test set with 7 : 3 (252 : 108), and the data in the training set were utilized to establish the risk model; firstly, the genes were verified as risk factors by univariate Cox regression analysis. Then, the genes with p < 0.05 were used to construct the multivariate Cox regression model, using the stepwise regression function step, with the parameter direction set to both, to adjust the multivariate regression model, and the obtained genes were used as prognostic factors to build the model. The risk value of each patient was counted by the expression of the genes, and the patients were grouped into high and low risk with the median risk value. Then, the risk profile was plotted and survival analysis for the high- and low-risk groups was conducted. In addition, we plotted the receiver operating characteristic (ROC) curve, and the area under curve (AUC) was used to indicate the prediction accuracy. Finally, the correlations between the risk model and clinical traits (age, gender, M, N, T, and other clinical data) were assessed using the chi-square test. Next, we validated the risk model using the TCGA test set and the ICGC validation set. In these two datasets, cases were spanided into high and low risks, respectively, and risk profiles, survival curves, and ROC plots were plotted, and correlations between risk factors and clinical traits were analyzed.
The clinical traits in the training set of TCGA-HCC data were collated, including age, sex, disease stage, T, N, and M. The samples were grouped according to the different clinical traits, and the risk values were compared between the different groups to see if there were significant differences and visualized by box plots.
The clinicopathological factors in the training set samples were added to the Cox analysis to investigate the independent prognosis of the risk model and clinicopathological factors. On this basis, a nomogram graph of the survival rate of the risk model and clinical factors was constructed. The factors that obtained significant results from the above multivariate Cox analysis were plotted, and the OS was predicted according to the total score. The correction curve was utilized to evaluate the prediction results of the model.
We divided the TCGA dataset into the high- and low-risk groups. The samples in the high- and low-risk groups were analyzed for differences using the “limma” R package, and the log2|FC| were then sorted from highest to lowest. Gene Set Enrichment Analysis (GSEA) was conducted using the “clusterProfiler” R package to find the common functions and related pathways of a large number of genes in the differentially expressed gene set [15]. The thresholds set were |NES| > 1, NOM p < 0.05, and q < 0.25, and the databases used for GSEA were Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO).
To further validate the accuracy of the risk model in predicting CAF, we executed Spearman correlation analysis on the risk score, stroma score, immune score, ESTIMATE score, tumor score, the proportion of CAF predicted by xCell, and the proportion of CAF predicted by EPIC, MCP-counter, and Tumor Immune Dysfunction and Exclusion (TIDE). Firstly, the “ESTIMATE” R package was utilized for ESTIMATE analysis to obtain the immune score, stromal score, ESTIMATE score, and tumor score for each sample. The EPIC algorithm analyzed the percentage of infiltration of eight-cell types, including CAFs, based on expression data [16]. We used the MCP-counter to attribute the content of CAFs in the samples. The xCell algorithm can also predict the proportion of CAFs. Finally, the CAF content was obtained using TIDE. The correlations between risk scores and each index were calculated using the Spearman correlation analysis. p < 0.05 represents significant correlation.
There were 23 CAF-associated marker genes, including ACTA2, ASPN, CAV1, COL11A1, COL1A1, COL1A2, COL3A1, EMILIN1, FAP, FN1, FOXF1, MFAP5, MMP11, MMP2, OGN, PDGFRA, PDGFRB, PDPN, S100A4, SLC16A4, SPARC, TNC, and ZEB1 [17, 18]. Then, we calculated the correlations between prognostic genes and risk scores with CAF marker genes.
ssGSEA is a single-sample GSEA method by which we can obtain the immune cell, of each sample [19]. Using 28 immune-related gene sets, we can get the immune activity. Then, the differences in 28 immune activities between the high- and low-risk groups were compared, and the differential immune activities were related to the risk scores.
We know that the Genomics of Drug Sensitivity in Cancer (GDSC) database has many drug sensitivity and genomic datasets that are important for the discovery of potential oncology therapeutic targets. IC50 refers to the half amount of a drug that inhibits specific biological processes. The “pRRopheticPredict” R package (version 0.5) was utilized to calculate 138 drugs included in the database and compare differences in drug IC50 between the high- and low-risk groups.
First, RNA was extracted from control cells WRL68 and HCC cells Huh7, Hepg2, and sk-sep-1, followed by a reverse transcription reaction, and finally, the target gene was amplified by PCR reaction. The RNA extraction kit was TRIzol Reagent (ref.: 15596018) kit provided by Ambion. The reverse transcription kit was the SweScript RT I First-strand cDNA Synthesis All-in-OneTM First-Strand cDNA Synthesis Kit (cat.: G33330-50) from Servicebio. PCR reactions were performed with the 2x Universal Blue SYBR Green qPCR Master Mix (cat.:G3326-05) kit from Servicebio. Primer sequences are shown in Table 1. The PCR reaction process was 95°C predenaturation for 1 min and then 40 cycles. Each cycle included 95°C denaturation for 20 s, 55°C annealing for 20 s, and 72°C extension for 30 s. The internal reference for gene detection is GAPDH. The expression of ACTA2, IGJ, CTHRC1, CXCL12, and LAMB1 in normal cell WRL68 and HCC cells Huh7, Hepg2, and sk-sep-1 were compared by analysis of variance (ANOVA), and p < 0.05 was a difference.
We calculated the immune cell content of 421 samples in the TCGA dataset (Figure 1(a)). After screening out the normal samples, there were 158 high CAF samples and 213 low CAF samples. The results of K-M analysis of the high and low CAF groups were shown (Figure 1(b)), and it can be seen that there was a significant survival difference between the high and low CAF groups. The results of clinical trait correlation between high and low CAF groups showed that CAF cells were different between different STAGE groups and between different GRADE groups (Figures 1(c) and 1(d)).
The clustering of the samples in the TCGA dataset was shown in Figure 2(a), and the samples were not deleted. The power threshold was chosen as 13, so that the interactions between genes conformed to the scale-free network (Figure 2(b)). From the module clustering tree, we can see that 12 modules were clustered, and after merging, 6 modules were obtained (Figure 2(c)). Finally, the key modules were filtered according to their correlation with CAF, and we got the green module (Figure 2(d)). Therefore, 898 genes in the green module were used as CAFGs.
There were 676 DEGs between the high and low CAF groups (Figure 3(a)). 6265 DEGs were found between normal and HCC samples (Figure 3(b)). CAFGs and DEGs between high and low CAF and DEGs between normal and HCC samples were crossed to obtain 107 CAF-DEGs, and the Venn diagram is shown (Figure 3(c), Table S1).
In the TCGA training set, univariate Cox analysis yielded 7 genes (Figure 4(a), Table 2). After multivariate Cox analysis, 5 genes appeared in multivariate Cox analysis (Figure 4(b), Table 3): ACTA2, IGJ, CTHRC1, CXCL12, and LAMB1. The risk value of each patient was counted from the expression of these five genes, and the cases were classified into high and low risks (median value = 0.988) (Figure 4(c)). The survival analysis of the high- and low-risk groups illustrated there was a significant survival difference between the high- and low-risk groups (Figure 4(d)). The AUC at 1, 3, and 5 years in the ROC curve were 0.661, 0.686, and 0.608, respectively (Figure 4(e)). In addition, in both the TCGA test set and ICGC validation set, the survival of the high-risk group was lower, and the AUC at 1, 3, and 5 years was more significant than 0.65 (Figures 5(a)–5(g)). In addition, in the ICGC validation set, grade was different between the high- and low-risk groups. It indicated that the risk model could be effectively used as a prognostic model.
The correlation between the risk model and clinical traits showed that the risk values differed significantly between stages I-II and stages III-IV. And the risk values were quite different between T1 − 2 and T3 − 4 stages. The results were shown (Figures 6(a)–6(f)).
The factors with p < 0.05 in the univariate Cox regression analysis were T, risk score, and stage (Figure 7(a), Table 4). The three significant factors were added to the multivariate Cox analysis (Figure 7(b), Table 5), and the results showed that risk score and stage were significant. The survival nomogram graph was shown (Figure 7(c)). In the corrected curve, the c-index was 0.703, and the corrected c-index was 0.696, and the slopes were calculated to be 0.697, 0.406, and 0.300 at 1, 3, and 5 years, which demonstrated the best prediction at one year (Figure 7(d)).
A total of 73 KEGG paths and 1968 GO paths were enriched by GSEA, and we selected the top 10 KEGG paths and GO paths to visualize them. As can be seen (Figure 8(a)), the top 10 KEGG pathways obtained have activation of the immune response, alcohol metabolic process, alpha-amino acid metabolic process, and B cell-mediated immunity. The top 10 GO functions were autoimmune thyroid disease, cell cycle, graft versus host disease, peroxisome, PPAR signaling pathway, and retinol metabolism (Figure 8(b)).
The correlation results of the risk score with other scores suggested that the risk score was negatively relevant to the immune score, ESTIMATE score, stromal score, xCell-predicted CAF ratio, and TIDE-predicted CAF ratio, and positively relevant with the tumor score (Figure 9(a)). The correlations between prognostic genes and risk scores with CAF marker genes were calculated, and the results were as follows. The correlation results illustrated that risk scores were negatively related toACTA2, ASPN, COL1A1, COL1A2, COL3A1, EMILIN1, FAP, FOXF1, MFAP5, MMP2, OGN, PDGFRA, PDPN, S100A4, SLC16A4, SPARC, and TNC genes. FN1 with LAMB1, CTHRC1, and SLC16A4 was positively associated with ACTA2, IGJ, CXCL12, and LAMB1. In addition, the remaining 21 CAF-related marker genes were positively associated with five prognostic genes (Figure 9(b)).
As can be seen (Figure 10(b)), among the 28 cells, 20 cells were different between the high- and low-risk groups, including activated B cell, CD56bright natural killer (NK) cell, CD56dim NK cell, central memory CD4 T cell, central memory CD8 T cell, and Type 1 T helper cell, and the 20 significant cells were plotted separately from the risk score in a lollipop plot as follows (Figure 10(a)).
According to the calculation results, 65 drugs showed differences in the high- and low-risk groups, which were temsirolimus, CI.1040, NU.7441, AZD8055, AICAR, AMG.706, DMOG, KU.55933, Metformin, EHT.1864, Dasatinib, NVP.BEZ235, PD.0325901, AZD.0530, NVP.TAE684, AKT.inhibitor.VIII, Vorinostat, GDC0941, PD.173074, Erlotinib, Docetaxel, WO2009093972, Rapamycin, AZD6244, JNJ.26854165, BI.D1870, MG.132, BX.795, A.770041, PD.0332991, Z.LLNle.CHO, AP.24534, Parthenolide, GW.441756, Nilotinib, OSI.906, X17.AAG, GDC.0449, AZD6482, WH.4.023, PF.4708671, Axitinib, TW.37, SB590885, Thapsigargin, NSC.87877, Cyclopamine, CMK, RDEA119, Gefitinib, Sorafenib, CEP.701, Imatinib, Methotrexate, ABT.263, Vinblastine, AZD7762, Lapatinib, AZ628, GNF.2, Bryostatin.1, Camptothecin, Nutlin.3a, FH535, and ZM.447439 (Table S2); they were visualized as a box plot as shown in the figure below. Figure 11 showed box plots for just the six drugs in the high- and low-risk groups.
The results of qPCR demonstrated that expression levels of ACTA2, IGJ, CTHRC1, CXCL12, and LAMB1 genes were different in normal cells WRL68 and HCC cells Huh7, Hepg2, and sk-sep-1. Specifically, ACTA2, CTHRC1, and LAMB1 genes were significantly upregulated in HCC cells Huh7, Hepg2, sk-sep-1, and IGJ, CXCL12 were downregulated in HCC cells (Figure 12).
While there have been advances in diagnostic techniques and treatment of HCC, [20, 21] the survival prognosis remains poor because of its high recurrence and metastasis rates [22]. CAFs are the main cellular component that can affect the formation of liver fibrosis, which in turn results in the development of HCC [10, 12]. Many prognostic models for HCC have been presented by far. Zhang et al. built a prognostic model which was able to reasonably predict the prognosis of HCC patients and provided a new idea to study HCC of different histological grades [21]. Long et al. developed a four-gene prognostic model to probe the differences in mRNA expression between HCC and neighboring liver to obtain potential genetic biomarkers [2]. Wang et al. screened immune-related differentially expressed genes closely related to HCC and further detected genes associated with prognosis [23]. However, because of the limitations of the public database data, further validation of the proposed prediction models is necessary or regression modeling methods need to be applied to determine if the prediction accuracy can be further improved. More than that, the validity of the prediction model should be confirmed in a large sample of HCC. In this study, we sought five biomarkers basing CAFGs for a prognostic model for HCC by bioinformatics method, conducted an independent prognostic analysis and functional enrichment analysis, and calculated the differences between immunoassay (immune infiltration, immunotherapy) and drug sensitivity at all levels. At last, qRT-PCR verified the expression levels of ACTA2, IGJ, CTHRC1, CXCL12, and LAMB1 genes in normal and HCC cells, which is a relatively complete work for the prognostic building. In the present study, five genes have been obtained for the HCC prognostic model. ACTA2, actin alpha 2, which contributed to cell-generated mechanical tension and maintenance of cell shape and movement, was highly expressed in carcinomas [24]. Meanwhile, a previous study showed that CAFs enhanced the tumor-initiating and tumorigenic properties of HCC cells, and ACTA2 was exactly a biomarker of CAFs. The upregulation of ACTA2 level indicated poor survival HCC patients [25]. It was demonstrated that a linking chain of multisomal IgA and IgM is also present in IGJ [26]. It is possible that their upregulation may enhance the anticancer immune response to sorafenib treatment and facilitate the survival of HCC [27, 28]. In addition, overexpression of CTHRC1 contributes to tumorigenesis and progression through positive regulation of tumor spread, invasion, migration, adhesion, and metastasis [29–31]. Immunohistochemical analysis demonstrated that CTHRC1 expression levels were elevated in HCC tissues [32]. Stromal-derived-factor-1 (SDF-1) was expressed in more than 23 different types and participated in tumor metastasis [33]. Interestingly, SDF-1 protein for the HCC cells was expressed in the cytoplasm and nucleus [34]. Notably, the level of SDF-1 was lower in HCC. Patients with relatively high SDF-1 showed longer OS [35]. LAMB1 consists of laminins [36]. LamB1 mediated β1 integrin signaling and can regulate cell migration, proliferation, and survival by activating specific p67kDa laminin receptors (LamR) [37–39]. HCC patients have shown elevated levels of LamB1 in cirrhotic tissues, with further increased expression in HCC [40]. In HCC, the expression of the b1 integrin receptor and LamR were upregulated, which was relevant with enhanced tumor aggressiveness and poor patient survival [41, 42]. Based on the enrichment analysis of the high- and low-risk groups by GSEA, function ways of fatty acid metabolism, amino acid metabolism, and immune response were related to the progress of HCC seriously. Firstly, a specific reprogramming xiang of fatty acid metabolism has been found in the nonalcoholic steatohepatitis (NASH) stage of nonalcoholic fatty liver disease (NAFLD). The liver is involved in the context of MetS and simple steatosis can progress to liver fibrosis or even cirrhosis, and eventually to HCC [43]. Metabolic reprogramming can support hepatocyte proliferation by participating in fatty acid synthesis and oxidation [44]. Second, the synthesis of nonessential amino acids is vital for the maintenance of liver function [45, 46]. In HCC, abnormalities in amino acid and protein metabolism occur [47]. Tumor immune cells can be participated in the immune response to cancer and also predict treatment efficacy and survival [48]. In the current study, there were 20 immune cells that differed between the high- and low-risk groups, including B cells, T cells, and NK cells. Regulatory B (Breg) cells accumulate in the tumor environment, and it can produce high levels of IL-10. Breg can suppress the host immune responses to promote tumorigenesis in HCC [49]. Regulatory T cells (Tregs), expressing CD25 and forkhead boxP3 (FoxP3), were negative during immune surveillance, resulting in tumor tolerance [50]. There are fewer NK cells in HCC tissue and NK cells can inhibit cytokine production and cytotoxic activity [51]. Zhu et al. constructed the prognostic model and the recurrence risk model and found that patients with high risk scores responded strongly to immune checkpoint inhibitor therapy and that low-risk patients may derive more significant clinical benefit from chemotherapy [52]. 65 drugs showed differences in the high- and low-risk groups. Temsirolimus is a prodrug of sirolimus. Studies have shown that temsirolimus has an inhibitory effect on HCC cells, and in phase I/II clinical trial, it was well-tolerated in HCC patients [53]. Moreover, temsirolimus is an mTOR inhibitor that can block cell cycle transition and affects cell proliferation by inhibiting mTOR and growth factors [54]. CI-1040, another drug predicted by our prognostic model, is an oral inhibitor of extracellular signal-regulated kinase (MEK) [55], It is a new candidate for targeted treatment of HCC because of its potential antitumor efficacy [56]. ZM447439 (ZM) induces apoptosis in HCC cells by interfering with spindle integrity and chromosome segregation [57]. These three drugs are representatives of anti-HCC drugs. However, among the 65 drugs, there are also some news, of which the effects on HCC are not definite. For example, GNF-2 inhibits the enzymatic and cellular kinase activities of ABL1, ABL2, and recombinant ABL and can inhibit the proliferation of fibroblasts. Still, its effect on anti-HCC have not been elucidated [58]. Then, AZ628, another new drug for HCC, can be involved in fibrosarcoma formation, and AstraZeneca can effectively inhibit cancer cell proliferation by inhibiting the activity of Raf [59]. CEP-701 can effectively inhibit trk receptors, leading to cell death in prostate cancer (PC), and it can also limit tissue penetration by binding serum proteins [60].
This study concentrated on the prognostic value of CAFs for HCC and identified CAF-related genes. A prognostic model of 5 CAFGs for HCC was developed in this research, and the expression of the five genes were verified by the qRT-PCR method. It provides new directions for the treatment of HCC. Nonetheless, one shortcoming of this study should be addressed, there are no clinical trials. | true | true | true |
||
PMC9590122 | Bing Wang,Li Yao,Yuefu Dong,Jian Liu,Jian Wu | LncRNA PCED1B-AS1 knockdown inhibits osteosarcoma via methylation-mediated miR-10a downregulation | 23-10-2022 | PCED1B-AS1,Osteosarcoma,miR-10a,Methylation,Proliferation | Background LncRNA PCED1B-AS1 (PCED1B-AS1) promotes glioma. This study aimed to investigate its role in osteosarcoma (OS). Methods The study included 60 OS patients. Accumulation of miR-10a and PCED1B-AS1 in tissues from OS patients and cell lines was determined by RT-qPCR. Cell transfections were performed for interaction analysis. Participation of PCED1B-AS1 siRNA silencing and miR-10a overexpression in proliferation, invasion, and migration of U2OS and MG-63 cells was analyzed by cell proliferation assay and Transwell assay. Results PCED1B-AS1 level was increased in OS and positively correlated with miR-10a level. In OS cells, PCED1B-AS1 siRNA silencing downregulated miR-10a. Methylation-specific PCR analysis showed that PCED1B-AS1 siRNA silencing decreased the methylation of miR-10a gene promoter. Moreover, PCED1B-AS1 siRNA silencing suppressed OS cell proliferation, invasion, and migration. In addition, miR-10a overexpression attenuated the effects of PCED1B-AS1 siRNA silencing. Conclusion PCED1B-AS1 knockdown may inhibit OS cell proliferation and movement by regulating miR-10 gene methylation. Supplementary Information The online version contains supplementary material available at 10.1186/s13018-022-03284-1. | LncRNA PCED1B-AS1 knockdown inhibits osteosarcoma via methylation-mediated miR-10a downregulation
LncRNA PCED1B-AS1 (PCED1B-AS1) promotes glioma. This study aimed to investigate its role in osteosarcoma (OS).
The study included 60 OS patients. Accumulation of miR-10a and PCED1B-AS1 in tissues from OS patients and cell lines was determined by RT-qPCR. Cell transfections were performed for interaction analysis. Participation of PCED1B-AS1 siRNA silencing and miR-10a overexpression in proliferation, invasion, and migration of U2OS and MG-63 cells was analyzed by cell proliferation assay and Transwell assay.
PCED1B-AS1 level was increased in OS and positively correlated with miR-10a level. In OS cells, PCED1B-AS1 siRNA silencing downregulated miR-10a. Methylation-specific PCR analysis showed that PCED1B-AS1 siRNA silencing decreased the methylation of miR-10a gene promoter. Moreover, PCED1B-AS1 siRNA silencing suppressed OS cell proliferation, invasion, and migration. In addition, miR-10a overexpression attenuated the effects of PCED1B-AS1 siRNA silencing.
PCED1B-AS1 knockdown may inhibit OS cell proliferation and movement by regulating miR-10 gene methylation.
The online version contains supplementary material available at 10.1186/s13018-022-03284-1.
As a type of primary malignant tumor originated from the skeleton, osteosarcoma (OS) is characterized by the formation of osteoid tissues and immature bones [1]. OS mainly affects teenagers and young adults [2], causing lifetime negative influence. It is estimated that OS accounts for more than 2% of malignancies in children blow 14 years old and 3% of malignancies in teens of 15 to 19 years old [1, 2]. OS diagnosed at early stages can be cured in most cases by neoadjuvant chemotherapy followed by surgical resection [3, 4]. However, tumor metastasis to other important organs, such as the lung and brain, is common in OS patients [5]. Distant tumor metastasis in OS patients is closely correlated with poor prognosis [6]. With localized tumors, more than 76% of OS patients can survive 5 years, while only about 25% of OS patients with distant metastasis can survive 5 years [5, 6]. Therefore, early diagnosis is the key to the survival of OS patients. However, due to lack of sensitive biomarkers, early diagnosis of OS is unlikely to be improved in the near future [3]. Therefore, in-depth investigation of the molecular pathogenesis of OS is still needed to improve the diagnosis and treatment of OS. Numerous previous studies on the molecular mechanism of OS have identified multiple molecular regulators in OS [7]. Functional characterization of these molecular players provides novel insights in OS management [8–10]. Non-coding RNAs (ncRNAs), such as microRNAs (miRNAs) and long ncRNAs (lncRNAs), are not involved in protein-coding but can play critical roles in human diseases, such as cancers, by directly or indirectly regulating gene expression [11, 12]. In fact, altered expression of ncRNAs, such as miRNAs, has shown promising potentials in the diagnosis and prognosis of cancers, and regulating the expression of miRNAs is an emerging novel therapeutic approach for cancer management [13]. Therefore, ncRNAs are the novel gold mine to identify potential targets for cancer targeted therapy [11, 12]. For instances, certain differentially expressed lncRNAs and miRNAs may be detected to diagnose cancers at early stages [11, 12]. Some lncRNAs with critical functions in regulating cancer cell behaviors may be regulated to suppress cancer progression [11, 12]. Moreover, lncRNAs may sponge miRNAs to suppress their activities and interact with other pathways, such as methylation pathways, to affect gene expression, thereby participating in cancers [11, 12]. It has been well established that miRNAs can affect DNA methylation by targeting methylation-related proteins and DNA methyltransferases to regulate the expression of lncRNAs [13]. More recently, lncRNAs are reported to affect m6A methylation, thereby regulating the expression of protein-coding genes and miRNAs [11, 12]. However, the function of most lncRNAs in cancers still has not been investigated, which limits their application in cancer diagnosis and treatment. LncRNA PCED1B-AS1 (PCED1B-AS1) promotes several types of cancers, such as glioma [14], pancreatic ductal adenocarcinoma [15], and hepatocellular carcinoma [16], while its role in other cancers is unknown. We performed preliminary deep sequencing analysis and observed the altered PCED1B-AS1 expression and its positive correlation with miR-10a (data not shown), a critical player in cancers [17]. In most cases, miR-10a promotes the development of different types of cancers via different cancer-related pathways. For instance, miRNA-10a is overexpressed in oral cancer and promotes glucose metabolism by upregulating GLUT1 to promote cancer cell proliferation [17]. Therefore, it is reasonable to hypothesize that PCED1B-AS1 may regulate cancer-related pathways through miR-10a to participate in OS. We then studied the crosstalk between PCED1B-AS1 and miR-10a in OS.
The study included 60 OS patients (37 males and 23 females; 12 to 33 years; 23.2 ± 3.4 years) who admitted to The First People's Hospital of Lianyungang between July 2016 and July 2019. The inclusion criteria were (1) OS patients diagnosed for the first time and (2) no therapy was initiated. The exclusion criteria were (1) patients complicated with other severe diseases, (2) patients with blood relationship and (3) recurrent cases. During biopsy, OS tissues and paired adjacent noncancerous tissues (normal bone tissues within 5 cm around tumors) were collected from each patient and freshly stored in liquid nitrogen before use. All tissue samples were confirmed by histopathological biopsy. The Ethics Committee of the aforementioned hospital approved this study (Ethical Approval No. #8631). Informed consent was obtained. Characteristics of patients are shown in Additional file 1: Table S1.
Human OS cell lines U2OS, MG-63, HOS, SJSA-1, and 143B (ATCC, USA) and normal osteoblast cells hFOB1.19 (ATCC, USA) were included in this study. MG-63, HOS, hFOB1.19 and 143B cell lines and SJSA-1 cells were cultured in EMEM (ATCC, USA) with 10% FBS and RPMI-1640 medium (ATCC, USA) with 10% FBS, respectively, at 37 °C (or at 33.5 °C for hFOB1.19 cells) in an incubator with 5% CO2 and 95% humidity.
PCED1B-AS1 siRNA (5′-AAGCGGUUCUCGUGCCUCAGU-3′), NC siRNA (Cat# SIC001, Sigma-Aldrich), miR-10a (5′-UACCCUGUAGAUCCGAAUUUGUG-3′) or NC miRNA (5′-GGUUCGUACGUACACUGUUCA-3′) was transfected into cells using Lipofectamine® 2000 (Invitrogen). In each transfection, 1 × 107 cells in a 10-cm dish were transfected with 50 nM miRNA and/or siRNA. NC siRNA- or NC miRNA-transfected cells were NC cells. Untransfected cells were control (C) cells. Subsequent experiments were performed at 48 h of post-transfection.
Total DNAs were extracted from U2OS and MG-63 cells using Quick-DNA Kit (ZYMO RESEARCH) and converted. The converted DNA samples were used as templates to perform MSP using 2X Taq FroggaMix (FroggaBio, USA) to analyze the methylation status of miR-10a gene with primers 5′-TTATTATTGTGTGTTCGGAAAATC-3′ (forward) and 5′-GTAACGCGCCTAACTATTTAACA-3′ (reverse) for methylated DNAs and 5′-TGTTTATTATTGTGTGTTTGGAAAATT-3′ (forward) and 5′-TCATAACACACCTAACTATTTAACAA-3′ (reverse) for un-methylated DNAs. PCR product was 1601 bp (from position -101 to -1701). PCR conditions were 5 min at 95 °C followed by 35 cycles of 95 °C for 30 s, 55 °C for 28 s and 72 °C for 35 s and a final extension at 72 °C for 10 min. All PCRs were conducted on a Bio-Rad C1000 PCR machine (Bio-Rad).
After RNA preparation using Direct-zol RNA, DNase I-digested RNA samples were used to prepare cDNA samples. With cDNA samples as templates, qPCRs were performed with 18S rRNA internal control to measure the accumulation levels of both PCED1B-AS1 and miR-10a. The method of 2−∆∆Ct was used for data normalizations. The following specific primers were employed: PCED1B-AS1 forward 5′-AAGGGGAAAGGAGGAAGTGAGAAG-3′ and reverse 5′-GGAAGCCAGTGAGCCAGGAGT-3′ and miR-10a forward 5′-CAGTGCAGGGTCCGAGGT-3′ and reverse 5′-GCCGTAC CCTGTAGATCCGAA-3′. PCR conditions were 1 min at 95 °C followed by 40 cycles of 95 °C for 10 s and 57 °C for 40 s. All qPCRs were conducted on BioRad CFX96 Touch Real Time PCR (Bio-Rad).
The proliferation of both U2OS and MG-63 cells after transfections was analyzed using a CCK-8 kit (Dojindo). Briefly, cells were harvested and cultured at 37 °C in a 96-well plate with 3000 cells in 0.1 ml medium per well. Three replicate wells were set for each experiment. Cell culture was performed for 48 h, followed by adding CCK-8 solution to 10%. After incubation with CCK-8 for 4 h, OD values at 450 nm were measured.
Transwell Inserts (8.0 μm, Corning) were used and cells in non-serum medium were added to the upper chamber. To induce cell movement, FBS was added to 20% in the lower chamber, and the upper chamber was filled with 6000 cells in serum-free media. After incubation at 37 °C for 24 h, cells on the lower membranes were stained with 1% crystal violet (Sigma-Aldrich) and counted.
Three independent replicates were included in each experiment, and mean ± SD values were used to express the data. Paired tissues (paired t test) and multiple groups (ANOVA Tukey’s test) were compared. The 60 OS patients were divided into high and low PCED1B-AS1 level groups (n = 30, cutoff value = median level of PCED1B-AS1 in OS tissues). Chi-squared test was applied to explore the associations between patients’ clinical characteristics and PCED1B-AS1 expression levels. Correlations were explored by performing Pearson’ correlation coefficient. p < 0.05 was statistically significant.
PCED1B-AS1 level was increased by 1.79-fold was in OS tissues compared to non-tumor tissue samples (Fig. 1A, p < 0.001). Chi-squared test was applied to explore the associations between patients’ clinical characteristics and PCED1B-AS1 expression levels. It was observed that PCED1B-AS1 expression was correlated with TNM stage, tumor metastasis, tumor size, and Enneking staging of OS, but not other factors (Additional file 1: Table S1, p < 0.05), suggesting the potential involvement of PCED1B-AS1 in the progression of OS. In addition, RT-qPCR analysis revealed that miR-10a expression levels were increased by 1.86-fold in OS tissues in comparison with the non-tumor tissues (Fig. 1B, p < 0.001).
Correlation analysis performed using Pearson’s correlation coefficient showed that PCED1B-AS1 expression levels were positively correlated with miR-10a expression levels across OS tissues (Fig. 2A, p < 0.0001), but not non-tumor tissues (Fig. 2B, p = 0.9935). RT-qPCR was also performed to analyze the expression of PCED1B-AS1 and miR-10a in OS cell lines and a normal cell line. As shown in Fig. 2C, D, both PCED1B-AS1 and miR-10a were upregulated in OS cells compared to the normal hFOB1.19 cell line, further confirming the upregulation of PCED1B-AS1 and miR-10a in OS.
Cell transfections were performed for interaction analysis. Considering that PCED1B-AS1 has already been accumulated to high levels in OS, PCED1B-AS1 siRNA silencing was performed. At 48 h after transfection, RT-qPCR analysis showed that PCED1B-AS1 was downregulated by 4.2-fold and miR-10a was upregulated by 5.1-fold (Fig. 3A, p < 0.05). RT-qPCR analysis also showed that PCED1B-AS1 siRNA transfection decreased miR-10a accumulation (Fig. 3B, p < 0.05). In contrast, miR-10a overexpression failed to significantly affect PCED1B-AS1 expression (Fig. 3B). Therefore, PCED1B-AS1 may serve as an upstream regulator of miR-10a in OS. To explore the potential mechanism, the role of PCED1B-AS1 in regulating miR-10a promoter region methylation was analyzed by MSP. MSP analysis showed that PCED1B-AS1 siRNA silencing increased miR-10a gene methylation (Fig. 3C).
The roles of PCED1B-AS1 siRNA silencing and miR-10a overexpression in the proliferation, invasion, and migration of U2OS and MG-63 cells were analyzed by cell proliferation (Fig. 4), invasion (Fig. 5A), and migration (Fig. 5B) assays. Compared with control cells without transfection, PCED1B-AS1 siRNA silencing decreased cell proliferation, invasion, and migration, and miR-10a overexpression increased cell proliferation, invasion, and migration (p < 0.05). In addition, miR-10a overexpression reversed the inhibitory effects of PCED1B-AS1 siRNA silencing on cell proliferation, invasion, and migration (p < 0.05) (Fig. 6).
We analyzed the interactions between PCED1B-AS1 and miR-10a in OS. PCED1B-AS1 and miR-10a were both upregulated in OS, and PCED1B-AS1 siRNA silencing decreased miR-10a methylation to suppress cell proliferation. It has been reported that PCED1B-AS1 plays a critical role in macrophage apoptosis and autophagy [18]. However, its functions in cancers have only been investigated in glioma [14]. PCED1B-AS1 is accumulated to high levels in glioma and regulates PCED1B through miR-194-5p to promote glioma [14]. Moreover, PCED1B-AS1 overexpression has been found to promote pancreatic ductal adenocarcinoma [15] and hepatocellular carcinoma [16]. We reported PCED1B-AS1 upregulation in OS. In addition, PCED1B-AS1 siRNA silencing reduced proliferation and movement of OS cells. Therefore, PCED1B-AS1 might play an oncogenic role in OS, and PCED1B-AS1 inhibition might serve as a potential target for the treatment of OS. However, clinical trials and animal model studies are needed to analyze the in vivo function of PCED1B-AS1 in OS and explore its potential clinical values. Different roles of miR-10a have been reported in different types of cancers [17, 19]. For instance, miR-10a was overexpressed in oral squamous cell carcinoma and promotes glucose metabolism in cancer cells by regulating glucose transporter 1 the expression [19]. In contrast, miR-10a is downregulated in colorectal cancer and suppresses epithelial-to-mesenchymal transition [19]. It has been reported that miR-10a is upregulated in OS [20], while its functions in OS remain unclear. Consistently, our study observed miR-10a upregulation in OS and its enhancing effects on OS cell proliferation, invasion, and migration. Therefore, miR-10a might play an oncogenic role in OS by promoting cancer cell proliferation. Glaich et al. reported that DNA methylation directly affects miRNA biogenesis. It is unknown whether lncRNA could regulate miRNA methylation. The key finding of the present study is that PCED1B-AS1 silencing downregulates miR-10a via methylation. However, methylation factors involved in this process remain to be further analyzed. Previous studies have shown that lncRNAs may interact with DNA methyltransferase [21]. For instance, HOTAIR upregulates DNA methyltransferases in hepatocellular carcinoma to epigenetically suppressed miR-122 [21]. In another study, PVT1 could recruit DNMT1 through EZH2 to miR-18b-5p gene promoter, thereby suppressing gene expression through methylation [22]. Future studies may focus on the potential interaction between PCED1B-AS1 and these methylation factors. It is unknown whether PCED1B-AS1 directly interacts with methylation factors to regulate miR-10a RNA gene via methylation or other mediators. In addition, we only observed the positive correlation between PCED1B-AS1 and miR-10a across OS tissue samples, but not non-tumor tissue samples. Therefore, the interaction between PCED1B-AS1 and miR-10a is likely mediated by certain pathological factors. Our data illustrated that PCED1B-AS1 silencing is likely a promising target to treat OS by negatively regulating multiple cancer cell behaviors. However, this study failed to analyze the diagnostic and prognostic values of PCED1B-AS1 for OS, especially its potential role in the early diagnosis of OS. Moreover, no in vivo experiment was performed to validate the interaction between PCED1B-AS1 and miR-10a. Future studies are still needed. With the increased understanding of the roles of non-coding RNAs in musculoskeletal conditions, novel diagnostic biomarkers and therapeutic approaches are expected to be developed [23–26].
PCED1B-AS1 and miR-10a are both upregulated in OS. PCED1B-AS1 siRNA silencing might serve as a potential target for the treatment of OS by suppressing OS cell proliferation. The function of PCED1B-AS1 in OS is likely mediated by regulating miR-10a through methylation.
Additional file 1. Supplemental Table 1: clinicopathologic characteristics of patient samples in OS. Supplemental Table 2: Correlation between PCEDB1-AS1 expression and clinicopathologic features in OS patients. | true | true | false |
|
PMC9590122 | Bing Wang,Li Yao,Yuefu Dong,Jian Liu,Jian Wu | LncRNA PCED1B-AS1 knockdown inhibits osteosarcoma via methylation-mediated miR-10a downregulation | 23-10-2022 | PCED1B-AS1,Osteosarcoma,miR-10a,Methylation,Proliferation | Background LncRNA PCED1B-AS1 (PCED1B-AS1) promotes glioma. This study aimed to investigate its role in osteosarcoma (OS). Methods The study included 60 OS patients. Accumulation of miR-10a and PCED1B-AS1 in tissues from OS patients and cell lines was determined by RT-qPCR. Cell transfections were performed for interaction analysis. Participation of PCED1B-AS1 siRNA silencing and miR-10a overexpression in proliferation, invasion, and migration of U2OS and MG-63 cells was analyzed by cell proliferation assay and Transwell assay. Results PCED1B-AS1 level was increased in OS and positively correlated with miR-10a level. In OS cells, PCED1B-AS1 siRNA silencing downregulated miR-10a. Methylation-specific PCR analysis showed that PCED1B-AS1 siRNA silencing decreased the methylation of miR-10a gene promoter. Moreover, PCED1B-AS1 siRNA silencing suppressed OS cell proliferation, invasion, and migration. In addition, miR-10a overexpression attenuated the effects of PCED1B-AS1 siRNA silencing. Conclusion PCED1B-AS1 knockdown may inhibit OS cell proliferation and movement by regulating miR-10 gene methylation. Supplementary Information The online version contains supplementary material available at 10.1186/s13018-022-03284-1. | LncRNA PCED1B-AS1 knockdown inhibits osteosarcoma via methylation-mediated miR-10a downregulation
LncRNA PCED1B-AS1 (PCED1B-AS1) promotes glioma. This study aimed to investigate its role in osteosarcoma (OS).
The study included 60 OS patients. Accumulation of miR-10a and PCED1B-AS1 in tissues from OS patients and cell lines was determined by RT-qPCR. Cell transfections were performed for interaction analysis. Participation of PCED1B-AS1 siRNA silencing and miR-10a overexpression in proliferation, invasion, and migration of U2OS and MG-63 cells was analyzed by cell proliferation assay and Transwell assay.
PCED1B-AS1 level was increased in OS and positively correlated with miR-10a level. In OS cells, PCED1B-AS1 siRNA silencing downregulated miR-10a. Methylation-specific PCR analysis showed that PCED1B-AS1 siRNA silencing decreased the methylation of miR-10a gene promoter. Moreover, PCED1B-AS1 siRNA silencing suppressed OS cell proliferation, invasion, and migration. In addition, miR-10a overexpression attenuated the effects of PCED1B-AS1 siRNA silencing.
PCED1B-AS1 knockdown may inhibit OS cell proliferation and movement by regulating miR-10 gene methylation.
The online version contains supplementary material available at 10.1186/s13018-022-03284-1.
As a type of primary malignant tumor originated from the skeleton, osteosarcoma (OS) is characterized by the formation of osteoid tissues and immature bones [1]. OS mainly affects teenagers and young adults [2], causing lifetime negative influence. It is estimated that OS accounts for more than 2% of malignancies in children blow 14 years old and 3% of malignancies in teens of 15 to 19 years old [1, 2]. OS diagnosed at early stages can be cured in most cases by neoadjuvant chemotherapy followed by surgical resection [3, 4]. However, tumor metastasis to other important organs, such as the lung and brain, is common in OS patients [5]. Distant tumor metastasis in OS patients is closely correlated with poor prognosis [6]. With localized tumors, more than 76% of OS patients can survive 5 years, while only about 25% of OS patients with distant metastasis can survive 5 years [5, 6]. Therefore, early diagnosis is the key to the survival of OS patients. However, due to lack of sensitive biomarkers, early diagnosis of OS is unlikely to be improved in the near future [3]. Therefore, in-depth investigation of the molecular pathogenesis of OS is still needed to improve the diagnosis and treatment of OS. Numerous previous studies on the molecular mechanism of OS have identified multiple molecular regulators in OS [7]. Functional characterization of these molecular players provides novel insights in OS management [8–10]. Non-coding RNAs (ncRNAs), such as microRNAs (miRNAs) and long ncRNAs (lncRNAs), are not involved in protein-coding but can play critical roles in human diseases, such as cancers, by directly or indirectly regulating gene expression [11, 12]. In fact, altered expression of ncRNAs, such as miRNAs, has shown promising potentials in the diagnosis and prognosis of cancers, and regulating the expression of miRNAs is an emerging novel therapeutic approach for cancer management [13]. Therefore, ncRNAs are the novel gold mine to identify potential targets for cancer targeted therapy [11, 12]. For instances, certain differentially expressed lncRNAs and miRNAs may be detected to diagnose cancers at early stages [11, 12]. Some lncRNAs with critical functions in regulating cancer cell behaviors may be regulated to suppress cancer progression [11, 12]. Moreover, lncRNAs may sponge miRNAs to suppress their activities and interact with other pathways, such as methylation pathways, to affect gene expression, thereby participating in cancers [11, 12]. It has been well established that miRNAs can affect DNA methylation by targeting methylation-related proteins and DNA methyltransferases to regulate the expression of lncRNAs [13]. More recently, lncRNAs are reported to affect m6A methylation, thereby regulating the expression of protein-coding genes and miRNAs [11, 12]. However, the function of most lncRNAs in cancers still has not been investigated, which limits their application in cancer diagnosis and treatment. LncRNA PCED1B-AS1 (PCED1B-AS1) promotes several types of cancers, such as glioma [14], pancreatic ductal adenocarcinoma [15], and hepatocellular carcinoma [16], while its role in other cancers is unknown. We performed preliminary deep sequencing analysis and observed the altered PCED1B-AS1 expression and its positive correlation with miR-10a (data not shown), a critical player in cancers [17]. In most cases, miR-10a promotes the development of different types of cancers via different cancer-related pathways. For instance, miRNA-10a is overexpressed in oral cancer and promotes glucose metabolism by upregulating GLUT1 to promote cancer cell proliferation [17]. Therefore, it is reasonable to hypothesize that PCED1B-AS1 may regulate cancer-related pathways through miR-10a to participate in OS. We then studied the crosstalk between PCED1B-AS1 and miR-10a in OS.
The study included 60 OS patients (37 males and 23 females; 12 to 33 years; 23.2 ± 3.4 years) who admitted to The First People's Hospital of Lianyungang between July 2016 and July 2019. The inclusion criteria were (1) OS patients diagnosed for the first time and (2) no therapy was initiated. The exclusion criteria were (1) patients complicated with other severe diseases, (2) patients with blood relationship and (3) recurrent cases. During biopsy, OS tissues and paired adjacent noncancerous tissues (normal bone tissues within 5 cm around tumors) were collected from each patient and freshly stored in liquid nitrogen before use. All tissue samples were confirmed by histopathological biopsy. The Ethics Committee of the aforementioned hospital approved this study (Ethical Approval No. #8631). Informed consent was obtained. Characteristics of patients are shown in Additional file 1: Table S1.
Human OS cell lines U2OS, MG-63, HOS, SJSA-1, and 143B (ATCC, USA) and normal osteoblast cells hFOB1.19 (ATCC, USA) were included in this study. MG-63, HOS, hFOB1.19 and 143B cell lines and SJSA-1 cells were cultured in EMEM (ATCC, USA) with 10% FBS and RPMI-1640 medium (ATCC, USA) with 10% FBS, respectively, at 37 °C (or at 33.5 °C for hFOB1.19 cells) in an incubator with 5% CO2 and 95% humidity.
PCED1B-AS1 siRNA (5′-AAGCGGUUCUCGUGCCUCAGU-3′), NC siRNA (Cat# SIC001, Sigma-Aldrich), miR-10a (5′-UACCCUGUAGAUCCGAAUUUGUG-3′) or NC miRNA (5′-GGUUCGUACGUACACUGUUCA-3′) was transfected into cells using Lipofectamine® 2000 (Invitrogen). In each transfection, 1 × 107 cells in a 10-cm dish were transfected with 50 nM miRNA and/or siRNA. NC siRNA- or NC miRNA-transfected cells were NC cells. Untransfected cells were control (C) cells. Subsequent experiments were performed at 48 h of post-transfection.
Total DNAs were extracted from U2OS and MG-63 cells using Quick-DNA Kit (ZYMO RESEARCH) and converted. The converted DNA samples were used as templates to perform MSP using 2X Taq FroggaMix (FroggaBio, USA) to analyze the methylation status of miR-10a gene with primers 5′-TTATTATTGTGTGTTCGGAAAATC-3′ (forward) and 5′-GTAACGCGCCTAACTATTTAACA-3′ (reverse) for methylated DNAs and 5′-TGTTTATTATTGTGTGTTTGGAAAATT-3′ (forward) and 5′-TCATAACACACCTAACTATTTAACAA-3′ (reverse) for un-methylated DNAs. PCR product was 1601 bp (from position -101 to -1701). PCR conditions were 5 min at 95 °C followed by 35 cycles of 95 °C for 30 s, 55 °C for 28 s and 72 °C for 35 s and a final extension at 72 °C for 10 min. All PCRs were conducted on a Bio-Rad C1000 PCR machine (Bio-Rad).
After RNA preparation using Direct-zol RNA, DNase I-digested RNA samples were used to prepare cDNA samples. With cDNA samples as templates, qPCRs were performed with 18S rRNA internal control to measure the accumulation levels of both PCED1B-AS1 and miR-10a. The method of 2−∆∆Ct was used for data normalizations. The following specific primers were employed: PCED1B-AS1 forward 5′-AAGGGGAAAGGAGGAAGTGAGAAG-3′ and reverse 5′-GGAAGCCAGTGAGCCAGGAGT-3′ and miR-10a forward 5′-CAGTGCAGGGTCCGAGGT-3′ and reverse 5′-GCCGTAC CCTGTAGATCCGAA-3′. PCR conditions were 1 min at 95 °C followed by 40 cycles of 95 °C for 10 s and 57 °C for 40 s. All qPCRs were conducted on BioRad CFX96 Touch Real Time PCR (Bio-Rad).
The proliferation of both U2OS and MG-63 cells after transfections was analyzed using a CCK-8 kit (Dojindo). Briefly, cells were harvested and cultured at 37 °C in a 96-well plate with 3000 cells in 0.1 ml medium per well. Three replicate wells were set for each experiment. Cell culture was performed for 48 h, followed by adding CCK-8 solution to 10%. After incubation with CCK-8 for 4 h, OD values at 450 nm were measured.
Transwell Inserts (8.0 μm, Corning) were used and cells in non-serum medium were added to the upper chamber. To induce cell movement, FBS was added to 20% in the lower chamber, and the upper chamber was filled with 6000 cells in serum-free media. After incubation at 37 °C for 24 h, cells on the lower membranes were stained with 1% crystal violet (Sigma-Aldrich) and counted.
Three independent replicates were included in each experiment, and mean ± SD values were used to express the data. Paired tissues (paired t test) and multiple groups (ANOVA Tukey’s test) were compared. The 60 OS patients were divided into high and low PCED1B-AS1 level groups (n = 30, cutoff value = median level of PCED1B-AS1 in OS tissues). Chi-squared test was applied to explore the associations between patients’ clinical characteristics and PCED1B-AS1 expression levels. Correlations were explored by performing Pearson’ correlation coefficient. p < 0.05 was statistically significant.
PCED1B-AS1 level was increased by 1.79-fold was in OS tissues compared to non-tumor tissue samples (Fig. 1A, p < 0.001). Chi-squared test was applied to explore the associations between patients’ clinical characteristics and PCED1B-AS1 expression levels. It was observed that PCED1B-AS1 expression was correlated with TNM stage, tumor metastasis, tumor size, and Enneking staging of OS, but not other factors (Additional file 1: Table S1, p < 0.05), suggesting the potential involvement of PCED1B-AS1 in the progression of OS. In addition, RT-qPCR analysis revealed that miR-10a expression levels were increased by 1.86-fold in OS tissues in comparison with the non-tumor tissues (Fig. 1B, p < 0.001).
Correlation analysis performed using Pearson’s correlation coefficient showed that PCED1B-AS1 expression levels were positively correlated with miR-10a expression levels across OS tissues (Fig. 2A, p < 0.0001), but not non-tumor tissues (Fig. 2B, p = 0.9935). RT-qPCR was also performed to analyze the expression of PCED1B-AS1 and miR-10a in OS cell lines and a normal cell line. As shown in Fig. 2C, D, both PCED1B-AS1 and miR-10a were upregulated in OS cells compared to the normal hFOB1.19 cell line, further confirming the upregulation of PCED1B-AS1 and miR-10a in OS.
Cell transfections were performed for interaction analysis. Considering that PCED1B-AS1 has already been accumulated to high levels in OS, PCED1B-AS1 siRNA silencing was performed. At 48 h after transfection, RT-qPCR analysis showed that PCED1B-AS1 was downregulated by 4.2-fold and miR-10a was upregulated by 5.1-fold (Fig. 3A, p < 0.05). RT-qPCR analysis also showed that PCED1B-AS1 siRNA transfection decreased miR-10a accumulation (Fig. 3B, p < 0.05). In contrast, miR-10a overexpression failed to significantly affect PCED1B-AS1 expression (Fig. 3B). Therefore, PCED1B-AS1 may serve as an upstream regulator of miR-10a in OS. To explore the potential mechanism, the role of PCED1B-AS1 in regulating miR-10a promoter region methylation was analyzed by MSP. MSP analysis showed that PCED1B-AS1 siRNA silencing increased miR-10a gene methylation (Fig. 3C).
The roles of PCED1B-AS1 siRNA silencing and miR-10a overexpression in the proliferation, invasion, and migration of U2OS and MG-63 cells were analyzed by cell proliferation (Fig. 4), invasion (Fig. 5A), and migration (Fig. 5B) assays. Compared with control cells without transfection, PCED1B-AS1 siRNA silencing decreased cell proliferation, invasion, and migration, and miR-10a overexpression increased cell proliferation, invasion, and migration (p < 0.05). In addition, miR-10a overexpression reversed the inhibitory effects of PCED1B-AS1 siRNA silencing on cell proliferation, invasion, and migration (p < 0.05) (Fig. 6).
We analyzed the interactions between PCED1B-AS1 and miR-10a in OS. PCED1B-AS1 and miR-10a were both upregulated in OS, and PCED1B-AS1 siRNA silencing decreased miR-10a methylation to suppress cell proliferation. It has been reported that PCED1B-AS1 plays a critical role in macrophage apoptosis and autophagy [18]. However, its functions in cancers have only been investigated in glioma [14]. PCED1B-AS1 is accumulated to high levels in glioma and regulates PCED1B through miR-194-5p to promote glioma [14]. Moreover, PCED1B-AS1 overexpression has been found to promote pancreatic ductal adenocarcinoma [15] and hepatocellular carcinoma [16]. We reported PCED1B-AS1 upregulation in OS. In addition, PCED1B-AS1 siRNA silencing reduced proliferation and movement of OS cells. Therefore, PCED1B-AS1 might play an oncogenic role in OS, and PCED1B-AS1 inhibition might serve as a potential target for the treatment of OS. However, clinical trials and animal model studies are needed to analyze the in vivo function of PCED1B-AS1 in OS and explore its potential clinical values. Different roles of miR-10a have been reported in different types of cancers [17, 19]. For instance, miR-10a was overexpressed in oral squamous cell carcinoma and promotes glucose metabolism in cancer cells by regulating glucose transporter 1 the expression [19]. In contrast, miR-10a is downregulated in colorectal cancer and suppresses epithelial-to-mesenchymal transition [19]. It has been reported that miR-10a is upregulated in OS [20], while its functions in OS remain unclear. Consistently, our study observed miR-10a upregulation in OS and its enhancing effects on OS cell proliferation, invasion, and migration. Therefore, miR-10a might play an oncogenic role in OS by promoting cancer cell proliferation. Glaich et al. reported that DNA methylation directly affects miRNA biogenesis. It is unknown whether lncRNA could regulate miRNA methylation. The key finding of the present study is that PCED1B-AS1 silencing downregulates miR-10a via methylation. However, methylation factors involved in this process remain to be further analyzed. Previous studies have shown that lncRNAs may interact with DNA methyltransferase [21]. For instance, HOTAIR upregulates DNA methyltransferases in hepatocellular carcinoma to epigenetically suppressed miR-122 [21]. In another study, PVT1 could recruit DNMT1 through EZH2 to miR-18b-5p gene promoter, thereby suppressing gene expression through methylation [22]. Future studies may focus on the potential interaction between PCED1B-AS1 and these methylation factors. It is unknown whether PCED1B-AS1 directly interacts with methylation factors to regulate miR-10a RNA gene via methylation or other mediators. In addition, we only observed the positive correlation between PCED1B-AS1 and miR-10a across OS tissue samples, but not non-tumor tissue samples. Therefore, the interaction between PCED1B-AS1 and miR-10a is likely mediated by certain pathological factors. Our data illustrated that PCED1B-AS1 silencing is likely a promising target to treat OS by negatively regulating multiple cancer cell behaviors. However, this study failed to analyze the diagnostic and prognostic values of PCED1B-AS1 for OS, especially its potential role in the early diagnosis of OS. Moreover, no in vivo experiment was performed to validate the interaction between PCED1B-AS1 and miR-10a. Future studies are still needed. With the increased understanding of the roles of non-coding RNAs in musculoskeletal conditions, novel diagnostic biomarkers and therapeutic approaches are expected to be developed [23–26].
PCED1B-AS1 and miR-10a are both upregulated in OS. PCED1B-AS1 siRNA silencing might serve as a potential target for the treatment of OS by suppressing OS cell proliferation. The function of PCED1B-AS1 in OS is likely mediated by regulating miR-10a through methylation.
Additional file 1. Supplemental Table 1: clinicopathologic characteristics of patient samples in OS. Supplemental Table 2: Correlation between PCEDB1-AS1 expression and clinicopathologic features in OS patients. | true | true | true |
|
PMC9590400 | Zhenguang Ying,Kaifang Wang,Junfeng Wu,Mingyu Wang,Jing Yang,Xia Wang,Guowei Zhou,Haibin Chen,Hongwu Xu,Stephen Cho Wing Sze,Feng Gao,Chunman Li,Ou Sha | CCHCR1-astrin interaction promotes centriole duplication through recruitment of CEP72 | 24-10-2022 | CCHCR1,Astrin,CEP72,Centrosome,Mitosis,Microtubule organization | Background The centrosome is one of the most important non-membranous organelles regulating microtubule organization and progression of cell mitosis. The coiled-coil alpha-helical rod protein 1 (CCHCR1, also known as HCR) gene is considered to be a psoriasis susceptibility gene, and the protein is suggested to be localized to the P-bodies and centrosomes in mammalian cells. However, the exact cellular function of HCR and its potential regulatory role in the centrosomes remain unexplored. Results We found that HCR interacts directly with astrin, a key factor in centrosome maturation and mitosis. Immunoprecipitation assays showed that the coiled-coil region present in the C-terminus of HCR and astrin respectively mediated the interaction between them. Astrin not only recruits HCR to the centrosome, but also protects HCR from ubiquitin-proteasome-mediated degradation. In addition, depletion of either HCR or astrin significantly reduced centrosome localization of CEP72 and subsequent MCPH proteins, including CEP152, CDK5RAP2, and CEP63. The absence of HCR also caused centriole duplication defects and mitotic errors, resulting in multipolar spindle formation, genomic instability, and DNA damage. Conclusion We conclude that HCR is localized and stabilized at the centrosome by directly binding to astrin. HCR are required for the centrosomal recruitment of MCPH proteins and centriolar duplication. Both HCR and astrin play key roles in keeping normal microtubule assembly and maintaining genomic stability. Supplementary Information The online version contains supplementary material available at 10.1186/s12915-022-01437-6. | CCHCR1-astrin interaction promotes centriole duplication through recruitment of CEP72
The centrosome is one of the most important non-membranous organelles regulating microtubule organization and progression of cell mitosis. The coiled-coil alpha-helical rod protein 1 (CCHCR1, also known as HCR) gene is considered to be a psoriasis susceptibility gene, and the protein is suggested to be localized to the P-bodies and centrosomes in mammalian cells. However, the exact cellular function of HCR and its potential regulatory role in the centrosomes remain unexplored.
We found that HCR interacts directly with astrin, a key factor in centrosome maturation and mitosis. Immunoprecipitation assays showed that the coiled-coil region present in the C-terminus of HCR and astrin respectively mediated the interaction between them. Astrin not only recruits HCR to the centrosome, but also protects HCR from ubiquitin-proteasome-mediated degradation. In addition, depletion of either HCR or astrin significantly reduced centrosome localization of CEP72 and subsequent MCPH proteins, including CEP152, CDK5RAP2, and CEP63. The absence of HCR also caused centriole duplication defects and mitotic errors, resulting in multipolar spindle formation, genomic instability, and DNA damage.
We conclude that HCR is localized and stabilized at the centrosome by directly binding to astrin. HCR are required for the centrosomal recruitment of MCPH proteins and centriolar duplication. Both HCR and astrin play key roles in keeping normal microtubule assembly and maintaining genomic stability.
The online version contains supplementary material available at 10.1186/s12915-022-01437-6.
Microtubules constitute an essential part of the cytoskeleton, maintaining cell shape and regulating mitosis [1]. During mitosis, microtubules extend from the centrioles, forming a spindle [2–4]. As the microtubular organization center, the centrosome is composed of a pair of centrioles and pericentriolar materials (PCM, also known as pericentriolar satellites) [5, 6]. Centrioles that display polar barrel-shaped structures with radial symmetry play a key role in the organization of centrosomes [6]. The number of centrioles in a cell is strictly regulated by the cell cycle. In the G1 phase, there is only one centrosome, which contains two isolated centrioles. PCM proteins are gradually recruited to the centrioles as the cell enters the S phase, and new procentrioles are formed at the proximal end of the existing centrioles. During the G2 phase, two centrosomes appear after duplication, and each contains two closely attached centrioles, which ensures that the daughter cells receive one centrosome with two centrioles after mitosis [7]. The PCM consists of various proteins, including pericentriolar materials 1 (PCM1), pericentrin, and a large number of centrosomal protein (CEP) family, such as CEP152, CEP63, and CEP215 (also named as cyclin-dependent kinase 5 regulatory subunit-associated protein 2 (CDK5RAP2)) [8]. These CEPs are not called a family in terms of homology, but they are all located in centrosomes, some of which are near the centriole and others are located in the outer part of the PCM, and perform different functions [9]. This complex structure of multiple, intertwined proteins is considered a platform for regulating organelle transport, spindle assembly, and cilia formation [10–12]. Astrin, a centrosome-related protein, which is also named sperm-associated antigen 5 (SPAG5) or mitotic spindle-associated protein p126 (MAP 126), dynamically localizes to the PCM, spindle poles, or outer kinetochores at different stages of the cell cycle. It participates in maintaining the dual-polarization of the spindle, the connection between microtubules and kinetochores, and the cohesion between sister chromatids, ensuring that mitosis proceeds properly. Deletion or mutation of astrin can lead to mitotic errors, such as spindle multi-polarization and chromosome separation failure [13–16]. In the centrosome, astrin is involved in the assembly of microcephaly (MCPH) proteins during interphase, which promotes centriole duplication [17]. The high expression of astrin is also positively correlated with the malignant degree of many tumors, indicating that its role in the centrosome is crucial [18–20]. Coiled-coil alpha-helical rod protein 1 (CCHCR1 or HCR) is a centrosome and processing body (P-body)-localized protein composed of multiple coiled-coil domains [21–23]. Although HCR has been widely reported as a susceptibility gene of psoriasis in genome-wide association studies, its function in cells is far from clear [24–27]. HCR interacts with them RNA-decapping protein 4 (EDC4) in the P-body, a special membraneless organelle dedicated to regulating mRNA decay and storage [23, 28–30]. However, the specific function of HCR in the P-body is unknown. HCR also exhibits a wide range of roles in various physiological processes, such as cell proliferation and steroid production [31, 32], and is also associated with alopecia areata, type-2 diabetes, and squamous cell carcinoma [33–35]. Interestingly, HCR has been predicted to interact with a series of centrosome- and mitosis-related proteins, such as PCM1, centrin, astrin, and CEP72, which suggests that HCR may participate in PCM networks and processes related to centrosome replication and mitosis [23]. In this study, we present evidence indicating that HCR is a key regulator of centrosome replication and microtubule organization. We show that HCR is localized and stabilized at the centrosome by directly binding to astrin. We also demonstrate that both HCR and astrin are required for the centrosome recruitment of CEP72 and MCPH proteins, including CEP152, CEP63, and CDK5RAP2. These findings provide a deeper understanding of the molecular function of HCR and are helpful for better exploring the role of HCR in psoriasis and other diseases.
In previous reports, exogenous HCR has been found to localize to the centrosomes and P-bodies, and several P-body- and centrosome-associated proteins have been identified as candidate interactors with HCR [23]. In this study, we also examined the binding partners of CCHCR1 by proximity-dependent biotinylation (BioID)-coupled mass spectrometry (LC-MS/MS). Similar to the data reported by Ling et al., we found astrin and mRNA-decapping protein 4 (EDC4) on the identified list (Table 1). Reciprocal immunoprecipitations were performed in HeLa cells to confirm the interaction between HCR and astrin. The endogenous immunoprecipitation experiments showed that astrin and HCR bound together as they were co-precipitated (Fig. 1A, Additional file 1: Fig. S1A). In 293 cells and U2OS cells, the exogenous and endogenous immunoprecipitation experiments performed showed the same results (Additional file 1: Fig. S1B). To further investigate whether there is a direct interaction between HCR and astrin, a GST pull-down assay was performed, and the results showed that HCR directly interacted with astrin in vitro (Fig. 1B). To map the binding sites between the two proteins, we analyzed the domains of HCR and astrin according to other studies [36, 37] and SMART Sequence Analysis Tools. Astrin consists of one unstructured region and two coiled-coil regions, whereas HCR contains three coiled-coil regions. Accordingly, we constructed a series of plasmids expressing truncated forms of HCR tagged with GFP or astrin tagged with myc (Fig. 1C). Immunoprecipitation assays revealed that the C-terminus of HCR (aa 441–782, coiled-coil 3, CC3) and the C-terminus of astrin (aa 893–1193, coiled-coil 2, CC2) mediated the interaction between them (Fig. 1D–F). In order to confirm whether there is a direct interaction in vitro, we also constructed a plasmid expressing GST-tagged astrin and a series of plasmids expressing truncated forms of His-tagged HCR to perform a GST pull-down experiment. The result confirmed that the third region of HCR interacted with astrin in vitro (Fig. 1G), which was consistent with the co-IP results in vivo. Previous studies have reported that astrin is located in the centrosome and spindle [13, 17]. To more precisely examine the intracellular localization of HCR, we generated a stable HeLa cell line transfected with GFP-tagged HCR (Additional file 2: Fig. S2A). Immunofluorescence (IF) staining showed that stably transfected HCR co-localized with astrin. In addition, the IF image of GFP-tagged-astrin-transfected HeLa cells co-stained with HCR and gamma-tubulin showed that astrin and HCR were co-localized in the centrosome (Fig. 2A). While the CC2 domain of astrin was sufficient to be recruited by the kinetochore [38], we also found that it was colocalized with the CC3 domain of HCR around centriolar (Additional file 2: Fig. S2B). This further confirms that HCR and astrin bind to each other through their C-terminus. In mitotic cells, HCR showed spindle localization indicated by alpha-tubulin, similar to that of astrin (Additional file 2: Fig. S2C). To confirm that the spindle localization of HCR is real and reliable, we also knocked down HCR by RNA interference (RNAi), and the results showed that the spindle localization of HCR disappeared (Fig. S2D). Also, GFP-tagged HCR showed co-localization with astrin throughout mitosis (Fig. 2B). Since both HCR and astrin co-immunoprecipitated with PCM1 (Fig. 2C), HeLa cells were stained with HCR and PCM1. The results showed that HCR only overlapped on the edges of the PCM1 throughout the cell cycle, except for telophase, which suggests that HCR may function as a bridge between the PCM and centriole (Fig. 2D). To further investigate whether HCR is also recruited to the centrosome via the microtubule transport system as PCM1, we disrupted the balance of microtubules using either the microtubule inhibitor nocodazole or microtubule stabilizer paclitaxel. Both treatments caused centrosome disintegration and disrupted the localization of HCR (Fig. 2E, Additional file 2: Fig. S2E), suggesting that the localization of HCR requires balanced microtubule dynamics. We also investigated whether HCR localization is regulated by PCM1 and pericentrin. Depletion of either PCM1 or pericentrin resulted in the delocalization of HCR from the whole centrosome (Fig. 2F), which indicated that the centrosome localization of HCR was controlled by both PCM1 and pericentrin. In turn, the knockdown of HCR did not affect PCM1 localization (Additional file 2: Fig. S2F). These results indicate that HCR is indeed a centrosome-associated protein and is under the control of the PCM platform.
To further analyze the functional relationship between HCR and astrin, we used siRNA to knockdown astrin and HCR in HeLa cells. Interestingly, depletion of astrin simultaneously reduced the protein level of HCR, while the protein level of astrin did not change after knockdown of HCR (Fig. 3A), and the decrease of HCR caused by depletion of astrin was not due to apoptosis or changes in the cell cycle (Additional file 3: Fig. S3A). Correspondingly, transient transfection of GFP-astrin in HeLa cells also increased the expression of endogenous HCR (Fig. 3B). These results suggested that astrin positively regulated the protein level of HCR. Additionally, IF staining showed that more HCR was recruited to the centrosome in cells overexpressing astrin as compared to astrin-depleted cells (Fig. 3C, D). By contrast, the depletion of HCR did not affect the centrosomal localization of astrin (Fig. 3E). To address the mechanism by which astrin affects the expression of HCR, we first examined whether astrin regulates HCR at the mRNA level. Real-time quantitative PCR results showed that the knockdown of astrin did not change the mRNA expression of HCR, suggesting that the regulation does not occur at the transcriptional level (Fig. 3F). Since the ubiquitin-proteasome pathway is one of the most common protein degradation pathways in mammalian cells [39], we speculated that astrin may affect the ubiquitination of HCR and then reduce the degradation of HCR. To test this hypothesis, an astrin knockout (KO) HeLa cell line was generated using the CRISPR/Cas9 technology and was verified by western blot (Additional file 3: Fig. S3B). It was shown that the level of HCR protein decreased significantly after astrin knockout. However, there was almost no difference in the expression level of HCR between astrin-KO and parental HeLa cells when treated with the proteasome inhibitor MG132 (Fig. 3G). Furthermore, immunoprecipitation analysis showed that the loss of astrin caused an increase in ubiquitinated HCR (Fig. 3H). Taken together, these results indicate that astrin protects HCR from ubiquitin-proteasome-mediated degradation and therefore maintains the protein level of HCR. Next, we questioned whether astrin was also responsible for the localization of HCR. The IF image in Fig. 3I showed that the recruitment of HCR on the centrosome was enhanced in HeLa cells treated with MG132. However, in astrin-KO cells treated with MG132, the centrosome localization of HCR did not significantly increase (Fig. 3I). Collectively, these results suggest that astrin not only protects HCR from ubiquitinated degradation, but also is responsible for the centrosome localization of HCR.
Another candidate binding partner of HCR is CEP72, a centrosome protein localized to the PCM [17, 40]. Both astrin and CEP72 are essential for the centrosome localization of a series of MCPH proteins, such as CDK5RAP2 (CEP215), CEP152, and CEP63, which ensure the successful duplication of centrioles [17]. Since astrin directly binds to CEP72, we wondered whether the association between HCR and CEP72 is direct or mediated by astrin or other proteins. Co-IP and GST pull-down assays confirmed that HCR directly binds to CEP72 with the third coiled-coil domain (Fig. 4A, B). As a cell cycle-dependent protein, the expression level of astrin changes at different stages of the cell cycle [15]. To examine the expression pattern of HCR and CEP72 in the cell cycle, HeLa cells at each cycle stage were obtained by the double-thymidine block method and analyzed by western blotting. It was revealed that the protein level of HCR increased from S to G2/M phase, peaked in the M phase, and then significantly decreased in the G1 phase, which was almost consistent with that of astrin, whereas the peak expression of CEP72 was later than that of astrin and HCR (Fig. 4C), suggesting that CEP72 might be under the regulation of astrin and HCR. In order to better understand whether astrin and HCR regulate CEP72, an HCR-knockout (KO) HeLa cell line was generated using CRISPR/Cas9 technology and was verified by western blotting (Additional file 2: Fig. S2A). Knocking out either HCR or astrin significantly reduced the signal of CEP72 on the centrosomes (Fig. 4D), while the expression level of CEP72 was almost unaffected (Additional file 4: Fig. S4). On the other hand, the depletion of CEP72 by siRNA did not affect the signals of HCR and astrin on the centrosomes (Fig. 4E).
Previous studies have revealed that centriole duplication relies on the centrosome localization of MCPH-associated proteins and PCM proteins. Among them, depletion of astrin or CEP72 reduced the recruitment of MCPH proteins, such as CEP152 and CEP63, to the centrosome, resulting in the inability of the centriole to duplicate properly from two to four foci [17, 41]. We found that knocking down HCR, astrin, or CEP72 by using siRNA lowered the 4 centriole foci ratio (Fig. 5A) and reduced the signals of CEP152 and CEP63 on the centrosomes (Fig. 5B) while not affecting their protein levels (Fig. 5C). In turn, depletion of CEP152 and CEP63 by siRNA did not affect the localization and expression levels of HCR, astrin, or CEP72 (Fig. 5D, E). Furthermore, immunoprecipitation analysis showed that HCR had no direct interactions with CEP152 and CEP63 (Additional file 4: Fig. S5). In addition to CEP152 and CEP63, another MCPH protein closely related to astrin-CEP72 recruitment is CDK5RAP2, which is also responsible for ensuring the replication of the centrosome [17]. Consistent with the results of astrin and CEP72 in the work of Kodani et al., the depletion of HCR by siRNA also caused the delocalization of CDK5RAP2 (Fig. 5F). These results suggested that, like astrin, HCR is also a key factor determining the centrosome localization of MCPH protein.
One of the most important roles of the centrosome is to regulate microtubule dynamics. PCM proteins play a critical role in the recruitment and assembly of microtubules. A previous study showed that depletion of CEP72 affected the nucleation activity of the microtubules and therefore decreased microtubule regrowth [40]. Similar results were obtained after the depletion of astrin and HCR by siRNA in HeLa cells (Fig. 6A) and RPE cells (Additional file 4: Fig. S6). It was reported that the destruction of the microtubule organization center could increase the length of microtubule plus-end tracking protein EB1 along the microtubules, which represents a decrease in the polymerization speed of the MT plus ends [42]. Compared with mock-treated cells, depletion of HCR, astrin, and CEP72 by siRNA caused a longer staining length of EB1, indicating that the polymerization of microtubules was slowed down (Fig. 6B). Together, these results revealed that lack of any of these three proteins could lead to microtubule nucleation defects and abnormal localization of EB1. Since the interaction between HCR and CEP72 relied on the C-terminal coiled-coil of HCR (CC3), we transfected GFP-tagged HCR-CC3 into HeLa cells to observe the effect on microtubule organization. In IF images, overexpressed HCR-CC3 showed many large puncta all over the cytoplasm, and the endogenous CEP72 was captured into these puncta, thus losing centrosome localization (Fig. 6C). This phenomenon indicated that overexpressed HCR-CC3 functioned as a dominant-negative inhibitor of endogenous HCR activity. Moreover, the microtubule organization center was seriously disrupted in HCR-CC3-transfected cells, which was in strong contrast to the clear microtubule aster in the surrounding non-transfected cells (Fig. 6D). These results provided further evidence that HCR-dependent centrosome localization of CEP72 is essential for microtubule organization.
Apart from their roles in centrosome replication, depletion of astrin or CEP72 also led to mitotic spindle pole defects and mitotic arrest [15, 40]. Cell cycle analysis by flow cytometry showed that almost half of the HCR-KO cells remained in M phase, while almost all the parental HeLa cells returned from M phase to G1 phase (Fig. 7A). This indicated that the loss of HCR might also lead to mitotic spindle defects and mitosis progression arrest. In mitotic cells, depletion of HCR by siRNA also caused multipolar spindle formation, suggesting that the absence of HCR could prevent the normal assembly of spindles (Fig. 7B, Additional file 4: Fig. S7). Similar results were obtained when astrin or CEP72 was knocked down, which is consistent with previous studies (Fig. 7B, Additional file 4: Fig. S7) [16, 40]. During the assembly of mitotic spindles, securin, a negative regulator of separase, can inhibit the production of activated separase before the onset of anaphase, which maintained the integrity of the mitotic centrosomes [43–46]. We showed here that securin was significantly downregulated, and separase was upregulated in HCR-depleted mitotic cells, similar to that in the astrin-depleted or CEP72-depleted cells (Fig. 7C) [15, 16]. This means that the absence of HCR, astrin, and CEP72 can cause abnormal activation of separase, which in turn leads to the polar division of the spindle to form a multi-polarization structure. In addition, we found an increased ratio of micronuclei in HCR-depleted cells, which indicates frequent chromosome segregation errors (Fig. 7D). In line with this phenomenon, IF results showed that phosphorylation of the DNA damage checkpoint kinases ATM (Fig. 7E) and gamma-H2AX (Fig. 7F) was increased in HCR-depleted cells. Western blot analysis showed that phosphorylation of Chk2 was also increased in HCR-depleted cells (Fig. 7G). These results suggested that the depletion of HCR caused frequent mitotic errors, resulting in genomic instability and DNA damage response. Astrin is also thought to be related to tumorigenesis [47–49]. To address whether HCR is involved in it, a colony formation assay was conducted. It showed that the knockout of astrin or HCR significantly impeded the colony formation ability of HeLa cells (Fig. 7H). To further verify that HCR knockdown could lead to a decrease in tumor proliferation, we constructed a subcutaneous transplantation tumor model in athymic mice. Tumor size in mice transplanted with either astrin-KO or HCR-KO cells was significantly smaller than that of mice transplanted with parental HeLa cells (Fig. 7I). These data indicated that loss of HCR is associated with a decrease in tumor proliferation, which may be due to a mitosis defect and genomic instability caused by HCR deletion.
HCR was initially reported as a centrosome and P-body-related protein [23, 32]. However, little is known about its cellular function and how it localizes to the centrosome. In this study, we provided evidence that HCR acted as an important link in the centrosomal protein recruitment chain. In fact, a variety of centrosomal components assemble at the centrosome in a PCM1-dependent manner, including centrin, ninein, astrin, and CEP131 [10]. PCM1 may deliver these proteins to the centrosome via the dynein-dynactin motor system [10, 50]. HCR is undoubtedly one of them because either depolymerization of the microtubule system or knockdown of PCM1 made HCR lose centrosome localization. Like astrin, CEP72, and CEP131, HCR did co-immunoprecipitate with PCM1. However, we would like to emphasize that astrin may play a more important role in maintaining centrosome localization of HCR. A previous study reported protein interaction between astrin and CEP72 [17]. Here, we show that astrin, HCR, and CEP72 interact with each other. Further analysis showed that astrin is in the most upstream position, which is essential for the centrosome localization of HCR and CEP72. HCR is in the middle, which does not affect astrin localization, but is required for CEP72 centrosome recruitment, while CEP72 is at the most downstream, which does not affect the positioning of HCR and astrin. However, we found that astrin was essential for stabilizing HCR and CEP72, whereas HCR and CEP72 had no significant effect on the protein level of astrin. It is worth noting that Kodani et al. reported that astrin and CEP72 stabilize each other, which differs from our results. The potential of a centrosome to anchor microtubules requires the correct assembly of a subset of proteins. According to the recruitment chain described by Kodani et al., CDK5RAP2 is recruited to the centrosome by astrin and CEP72, followed by CEP152, WDR62, and CEP63 in a stepwise, hierarchical manner, and finally comes CDK2, a protein kinase critical for centriolar duplication [17]. The localization of HCR is in the middle of PCM1 and centrin1 (Additional file 5: Fig. S8), which means that it may act as part of the chain linking PCM and centriole. We did find that depletion of HCR phenocopied the effect of astrin or CEP72 depletion on the centrosomal localization of CDK5RAP2. Accordingly, the centrosomal localization of CEP152 and CEP63, two factors downstream of CEP72, were also regulated by HCR, but no direct interactions were detected (Additional file 4: Fig. S5). In addition, we found that there was an interaction between HCR and CEP131 (also named AZI1) (Additional file 5: Fig. S9), which is consistent with the predictions of Ling et al. [23]. In the study of Kodani et al., CEP131, as a pericentriolar satellite protein, was responsible for ensuring the localization of CEP152 [17]. The interaction between HCR and CEP131 suggests that the recruitment of these MCPH proteins to the centrosome is more complicated than currently known. Like HCR, CEP131 is also considered to play an important role in maintaining genomic stability and tumor proliferation [51, 52]. One of the important roles of astrin in mitosis is to strengthen the connection between microtubules and the outer kinetochore of the chromosome, allowing the chromosome to withstand the tension from the spindle filament. In this process, astrin forms a complex with SKAP, MYCBP, and LC8 in kinetochore microtubules [36, 37, 53]. However, our results did not support the interaction between HCR and this complex (Additional file 5: Fig. S10). Although there is no evidence that HCR localizes to the kinetochore, it is still possible that HCR indirectly influences the role of astrin at the kinetochore, such as the transport of astrin between the spindle pole and the kinetochore, just like NuMA does [54]. Interestingly, we also found an interaction between HCR and NuMA (Additional file 5: Fig. S11). There may be an unknown relationship between NuMA and HCR on the spindle, which can affect or be affected by astrin to participate in the assembly and activity of mitotic spindles. Alternatively, HCR may be associated with important kinases, such as Plk-1 or PP1, which are responsible for the phosphorylation of astrin on kinetochore [36, 55, 56]. Another important role of astrin is to participate in the cohesion between sister chromatids in mitosis, which is the key point at which the existence of astrin can prevent early activation of separase before the onset of anaphase [15, 16]. In this study, we found that the knockdown of HCR increased the expression level of the active form of separase in M phase cells (Fig. 6E), suggesting that HCR is likely to affect sister chromatid cohesion. These critical mitosis processes are regulated by Aurora, a key family of kinases in charge of mitosis [57–60]. Since the Aurora kinases regulate the active conversion of astrin during mitosis, it is definitely worth exploring whether they also regulate HCR [53, 61, 62]. HCR is also localized to P-bodies and interacts with EDC4. Astrin was reported to recruit raptor to stress granules (SGs) upon oxidative stress, where it colocalized with G3BP1, an SG marker [63]. In fact, P-bodies and SGs are closely linked in function [64]. Interestingly, we also found that HCR co-localized with astrin and EDC4 in HeLa cells treated with arsenite (Additional file 5: Fig. S12), and the centrosomal protein CEP85 was also considered to be related to P-bodies [65]. Furthermore, we also found that EDC4 co-localized with the HCR in the centrosome and punctate staining around the spindle during mitosis (Additional file 5: Fig. S13). Additionally, a pair of P-bodies were found to reside at the centrosome in U2OS cells, as well as diverse non-malignant cells [66, 67]. Although the mechanism is unknown, the knockdown of some P-body components by RNA interference impaired primary cilium formation in human astrocytes [67]. Further in-depth study of HCR may reveal clearer functional links between the two structures. In a more macroscopic direction, the elucidation of the intracellular mechanisms of HCR also contributes to the understanding of various diseases. Recent reports have proposed that HCR is closely related to alopecia areata, psoriasis, and diabetes [31, 33, 68]. HCR-deficient mice showed stress-induced alopecia [35]. Since primary cilia play an important role in the development of hair follicles, the role of HCR in ciliogenesis deserves future attention. In addition, the interaction between HCR and astrin also suggested that HCR might be related to cancers. Numerous reports have confirmed that astrin overexpression is often associated with malignancy, so HCR, as a protein regulated by astrin, may also be upregulated in tumor tissues [20, 48, 49, 53, 69, 70]. Moreover, analysis of the TCGA database (portal.gdc.cancer.gov) also revealed that the transcription of HCR did significantly increase in a variety of tumors (Additional file 5: Fig. S14). Although we did not investigate whether HCR was involved in tumorigenesis, as a regulator of the cell cycle and mitosis, this possibility exists. Interestingly, HCR may even have a potential link with COVID-19 [71]. In fact, linking microtubule and centrosome to virus infection is not a new idea. Previous studies have found that retroviruses, such as human immunodeficiency virus type 1 (HIV-1) infection, can affect changes in centrosome function [72]. Although there is no clear evidence as to whether HCR is a direct target of COVID-19, further in-depth research may help explain the biological mysteries of the centrosome and provide substantial clinical value.
In conclusion, our results reveal the role of previously unfocused P-body protein HCR on centrosome, whereby HCR interacts with astrin to recruit CEP72 and MCPH proteins to the centrosome and ensures efficient centriole replication and other centrosome-related functions such as spindle-pole formation and microtubules organization (Fig. 8). Therefore, HCR not only acts as P-body component, but also plays an important role in the development of centrosome and the stability of the genome.
Human CCHCR1 cDNA (NM_019052) was amplified from HeLa cDNA by PCR amplification and subcloned into pEGFPN1 or pmCherryN2 vectors. Human astrin cDNA (NM_006461) in the pEGFPC2 vector was gifted by Dr. Yi-Ren Hong (Kaohsiung Medical University, Taiwan China) [73] and was subcloned into the pCMV-myc vector. CEP72 cDNA (NM_018140) was amplified from the pEBTet-CEP72-SNAP plasmid purchased from Addgene (plasmid #136819) and subcloned into the pEGFPN1 vector. Serial deletion fragments of indicated regions of HCR and astrin were amplified from HCR and astrin cDNA, respectively, and subcloned into pEGFPN1 and pCMV-Myc vectors, respectively. Antibodies used in this study included: astrin (14726-1-AP, for western blotting (WB) 1:2000, for immunofluorescence (IF) 1:500); CEP72 (19928-1-AP, for WB 1:1000, for IF 1:400); CEP152 (21815-1-AP, for WB 1:1000, for IF 1:400); CEP63 (16268-1-AP, for WB 1:1000, for IF 1:400); CEP131 (25735-1-AP, for WB 1:1000); centrin-1 (12794-1-AP, for IF 1:400) from Proteintech (Wuhan, China); CCHCR1 (sc-135052, for IF: 1:100, WB 1:500); γ-tubulin (sc-17788, for IF 1:200); cyclin B1 (sc-245, for WB 1:500); cyclin E (sc-247, for WB 1:500); cyclin D1 (sc-246, for WB 1:500); securin (sc-56207, for WB 1:500); separase (sc-390314, for WB 1:500); EB1 (sc-47704, for IF 1:100); PCM1 (sc-398365, for IF 1:200, for WB 1:500); pericentrin (sc-376111, for WB 1:500. IF 1:200); c-myc (sc-40, WB 1:1000) from Santa Cruz Biotechnology (Dallas, Texas, USA); γ-tubulin (GTX113286, for IF 1:500); astrin (GTX115449, for IF 1:400, WB 1:1000) from Genetex (Irvine, CA, USA); CDK5RAP2 (A15476, for IF 1:200); GFP (AE012, WB1:1000); and mCherry (AE002, WB 1:1000) from Abclonal (Wuhan, China). Nocodazole (GC14075) and thymidine (GC15815) were purchased from GPLBIO (Montclair, CA, USA). Paclitaxel (S1748) and MG132 (SC0213) were purchased from Beyotime Biotechnology (Shanghai, China).
The cell lines of HeLa (human cervical carcinoma cell), U2OS (human osteosarcoma cells), hTERT-RPE1 (immortalized human retinal pigment epithelial cells), and HEK293 (human embryonic kidney cells) were obtained from China Center for Type Culture Collection (Wuhan, China) and were cultured in DMEM high-glucose medium (Hyclone, Waltham, MA, USA) with 10% fetal bovine serum (FBS, Gibco, Waltham, MA, USA), 100 units/ml penicillin, and 10 μg/ml streptomycin (Biosharp, Hefei, China) in a humidified chamber with 5% CO2 at 37 °C. HCR-KO HeLa cells and astrin-KO HeLa cells were customized by VigeneBio (Jinan, Shandong, China) and Ubigene (Guangzhou, Guangdong, China) and cultured under the same conditions as HeLa cells.
The HCR-KO cell line was created by using CRISPR-Cas9 in HeLa cells with sgRNAs as follows: sgRNA1: CCCGAATGGTGTGGACCTTG sgRNA2: GCGGGAAGAACGGAACCGCC sgRNA3: AACGGGATGTTTCCAGTGAC sgRNA4: TGAGGTTGTCCGGAAGAACT These sgRNAs are designed to target exon 3-13 of the human CCHCR1 gene. The astrin-KO cell line was created by using CRISPR-Cas9 in HeLa cells with sgRNAs as follows: SPAG5-gRNA1: CTCTACTCCTAAAACGTCTG AGG SPAG5-gRNA2: ACCAGATCGTCTGTTCTCAA AGG These sgRNAs are designed to target exon 3 of the human SPAG5 gene. The specific verification reports refer to Additional file 7 and Additional file 8.
HeLa cells and HCR-KO HeLa cells were first synchronized with 5 mM thymidine for 16 h, washed with phosphate-buffered saline (PBS) three times, and cultured in DMEM without thymidine for 12 h, After treatment with 5 mM thymidine for another 12 h, cells were released from thymidine and harvested at each time point according to experimental needs. For collecting mitotic cells, cells were released for about 10 h from a double-thymidine block to initiate prometaphase [54]. For separase and securin analysis in mitotic cells, cells were treated with siRNA for 72 h and incubated with nocodazole (100 ng/ml in medium) for another 16 h [15, 16].
HeLa cells were transfected with 15 μg of DNA plasmid in a 10-cm dish or 2 μg in each well of a 6-well plate using Lipo6000 Transfection Reagent (Beyotime Biotechnology, Shanghai, China) following the manufacturer’s instructions. Cells were harvested and then lysed for co-IP or fixed for IF after treatment for 24 h.
HeLa cells were transfected with 10 nM siRNA using Lipo6000 Transfection Reagent (Beyotime Biotechnology, Shanghai China) following the manufacturer’s instructions. The cells were harvested and then lysed or fixed for further analysis after treatment for 72 h. The CCHCR1 ON-TARGETplus SMARTpool siRNA was purchased from Dharmacon (Lafayette, CO, USA). The siRNAs targeting astrin (5′-CAAUACCAAGACCAACUGG-3′), CEP72 (5′-TTGCAGATCGCTGGACTTC-3′), CEP152 (5′-GCAUUGAGGUUGAGACUAA-3′), CEP63 (5′-GAGUUACAUCAGCGAGAUA-3′), Percentrin (5′-GCAGCUGAGCUGAAGGAGA-3′), and PCM1 (5′-UCAGCUUCGUGAUUCUCAG-3′) were synthesized by Ribobio (Guangzhou, Guangdong China).
For the immunoprecipitation, plated cells were washed three times with PBS and then lysed with RIPA buffer (50 mM Tris, 150 mM NaCl, 0.1% NP40, with cocktail protease inhibitors (MCE Monmouth Junction, NJ, USA, Cat. No.: HY-K0011)) for 30 min on ice. Samples were then centrifuged at 14,000 rpm for 30 min to obtain lysate, and 5% of the lysates were saved as input. Then, 500 μg of the lysates was incubated with the 2 μg of antibodies for 2 h at 4 °C on a rotator, and then 50 μl of a mixed suspension of 50% protein A and protein G beads (pre-washed with PBS 3 times) was then added. Mixtures were incubated at 4 °C for 16 h on a rotator. The beads were collected by centrifuging at 2000 rpm for 2 min at 4 °C and then washed with PBS 3 times. The samples were eluted by resuspending washed beads in 30–50 μl of 2× SDS-loading buffer and heating at 95 °C for 5 min, followed by separation via SDS-PAGE and immunoblotting with appropriate antibodies.
For immunofluorescence imaging, cells plated on glass coverslips were fixed with cold methanol, blocked with 10% FBS, and probed with primary antibodies and then secondary antibodies coupled with AlexFluor 488/555/594/647. DNA was stained with DAPI. Immunofluorescence pictures were imaged under an Olympus Confocal Laser Scanning Microscope FV3000 (Olympus Co., Tokyo, Japan) and processed by ImageJ (https://imagej.nih.gov/ij/download.html) when necessary.
siRNA-treated or plasmid-transfected cells were treated with 1 μM nocodazole on ice for 30 min to depolymerize the microtubules and were then released from cold nocodazole after 0 min and 5 min to repolymerize the microtubules. For the microtubule regrowth assay, the cells were fixed and co-stained with gamma-tubulin and alpha-tubulin to show the microtubule organization center and microtubules, and the length of microtubules of each cell was measured to compare the differences between the groups.
For cell cycle analysis, cells were trypsinized and fixed in 70% ethanol at 4 °C for 16 h, washed with PBS 3 times, and stained with 50 mg/ml propidium iodide (PI, DNA stain) and 0.025 mg/ml RNase A in PBS for 30 min at 37 °C. Cells were analyzed with FACS Calibur (Becton Dickinson, Franklin Lakes, NJ, USA). The cell cycle results were analyzed based on the DNA content in cells. For statistical analysis, the results of 10,000 cells in each group were counted and plotted.
Total RNA was isolated with TRIzol (Invitrogen, Waltham, MA, USA) and used for cDNA reverse transcription with the Goldenstar RT6 cDNA synthesis kit (Tsingke, Beijing, China). Quantitative PCR analysis of gene transcripts was performed by the qPCR method using qPCR Master Mix (Promega, Madison, WI, USA) and Jena qTOWER3 system with the expression of GAPDH as the endogenous control.
Parental HeLa cells, HCR-KO HeLa cells, and astrin-KO HeLa cells were maintained in culture media in a 10-cm dish for 2 weeks, followed by staining with Giemsa stain. Then the number of stained colonies were counted.
Animals were randomly grouped in three groups with 5 mice per group. Parental HeLa cells, HCR-KO HeLa cells, or astrin-KO HeLa cells were injected into the subcutaneous prothorax of 6-week-old athymic mice with 1 × 106 cells per mice (BALB/c, Guangzhou Medical Animal Center, Guangzhou, China). After visible tumors were observed, tumor size was measured every 3 days and calculated according to the following formula: length × width. The measurement and data processing were performed with blinding. All mice received a humane diet and living environment during the experiment. At the end of the experiment, all mice were executed in a humane manner, and the subcutaneous tumor was exfoliated and weighed. This study was approved by the Animal Care Committee of Shenzhen University Science Health Center.
For the construction of HCR fragments plasmids, the SMART Sequence Analysis Tools (https://smart.embl-heidelberg.de) was used to analysis the protein domains.
For western blot results and immunofluorescence images, ImageJ (https://imagej.nih.gov/ij/download.html) was used to measure the intensity of the protein of interest. Microsoft Office Excel and GraphPad Prism were used to perform statistical analyses and graphing. For statistical analysis of blotting experiments, each experiment was performed three times independently. For statistical analysis of immunofluorescence images, 100 cells were counted from three independent experiments. All statistical results are presented as mean ± SD and tested with a two-tailed Student’s t test (GraphPad Prism software) to calculate the P-values between unpaired samples. The differences were considered statistically significant when P < 0.05.
Additional file 1: Fig. S1. Repeated verification of the interaction between HCR and astrin. (A) Astrin Co-IP HCR and CEP72 and HCR Co-IP astrin and CEP72. HeLa cell lysates were immunoprecipitated with astrin, HCR, or control rabbit IgG antibodies and analyzed by western blotting with anti-astrin, anti-HCR antibodies. Anti-GM130 and anti-beta actin antibodies were used as negative control. The blotting of astrin, HCR and GM130 antibodies were incubated on the same membrane by repeatedly washing the membrane with antibody removal solution to increase comparability. (B) HCR interacts with astrin in HEK293 and U2OS cells. mCherry vector alone or HCR-mCherry in conjunction with the GFP-astrin plasmid were transfected into HEK293 cells for immunoprecipitation using an mCherry antibody. The eluted proteins were analyzed with mCherry and GFP antibodies. HCR-mCherry plasmid-transfected HEK293 cells were immunoprecipitated with astrin antibody or negative control rabbit IgG. The precipitates were analyzed with mCherry and astrin antibodies. U2OS cell lysates were immunoprecipitated with an HCR antibody or negative control rabbit IgG. The precipitates were analyzed with antibodies against HCR and astrin.Additional file 2: Fig. S2. Identification of HCR-GFP stable Cell Line and Localization of HCR in cells. (A) Identification of HCR-KO HeLa cell line and stably expressing HCR-GFP cell line. Parental HeLa cells, HCR-KO HeLa cells, HCR-KO cells transfected with HCR-GFP, and stably transfected HCR-GFP HCR-KO cells were immunoblotted with an HCR antibody. (B) Co-localization of astrin-CC2 and HCR-CC3. HeLa cells transfected with astrin-CC2-myc and HCR-GFP or astrin-CC2-myc and HCR-CC3-GFP were co-stained with myc (red) and gamma-tubulin (cyan); scale bars, 10 μm. (C) Co-localization of HCR with alpha-tubulin. Mitotic HeLa cells stained with an alpha-tubulin antibody (green), HCR antibody (red), and DAPI (blue) for nuclear staining (left panel) or stained with anti-alpha-tubulin (green), anti-astrin (red), and DAPI (blue) (right panel); scale bars, 10 μm. (D) Identification of antibody staining to HCR. Negative control, HCR siRNA-treated HeLa cells were co-stained with HCR (red) and alpha-tubulin (green); scale bars, 10 μm. (E) The effect of Nocodazole on HCR is dose-dependent and recoverable. HeLa cells were treated with 1μM, 0.75μM, 0.5μM Nocodazole for 5 hours or treated with 1μM Nocodazole for 5hours then released from Nocodazole for 30 min, 1 hour, 2 hours, then co-stained with HCR (red) and gamma-tubulin (green). (F) Knockdown of HCR does not affect PCM1 localization. Negative control, HCR siRNA-treated HeLa cells were co-stained with HCR (red) and PCM1 (green); scale bars, 10 μm.Additional file 3: Fig. S3. Apoptosis or cycle changes in astrin-KO cells and identification of astrin-KO cell line. (A) Parental, astrin-KO and HCR-KO HeLa cells were analyzed with astrin, HCR, cyclin B1 and Cleaved PARP antibodies. HeLa cells treated with DMSO, Paclitaxel, Bafilomycin and MG132 were analyzed with Cleaved PARP and HCR antibodies. (B) Parental and astrin-KO HeLa cells were immunoblotted with an astrin antibody.Additional file 4: Fig. S4. Loss of either HCR or astrin slightly affects the protein level of CEP72. Parental HeLa cells, HCR-KO HeLa cells, and astrin-KO HeLa cells were analyzed by immunoblotting with CEP72 and beta-actin antibodies. Fig. S5. HCR does not bind to CEP63 and CEP152. HeLa cell lysates were immunoprecipitated with antibodies specific for HCR and negative control IgG. The precipitates were analyzed by immunoblotting with antibodies against CEP63 and CEP152. Fig. S6. Knockdown of astrin and HCR also caused microtubules organization defects in RPE cells. Microtubule regrowth assay of negative control, astrin and HCR siRNA-treated RPE cells co-stained with gamma-tubulin (green), alpha-tubulin (red), and DAPI (blue); scale bars, 10 μm. Fig. S7. Knockdown of astrin and HCR also caused mitotic spindle defects in RPE cells. Negative control, astrin and HCR siRNA-treated HeLa cells were treated with 100 ng/ml nocodazole for 16 hours to be arrested in M-phase and co-stained with gamma-tubulin (green), alpha-tubulin (red), and DAPI (blue); scale bars, 10 μm.Additional file 5: Fig. S8. Co-localization of HCR with centrin1 and PCM1. HeLa cells transfected with HCR-GFP were co-stain with centrin1 (red) and gamma-tubulin (cyan) and DAPI (blue) or PCM1 (red) and gamma-tubulin (cyan) and DAPI (blue); scale bars, 10 μm. Fig. S9. HCR interacts with CEP131. HeLa cell lysates were immunoprecipitated with antibodies specific for HCR, astrin, or negative control rabbit IgG. The precipitates were analyzed by immunoblotting with antibodies against HCR and CEP131. Fig. S10. HCR did not interact with SKAP, MYCBP and LC8. HeLa cell lysates were immunoprecipitated with astrin, HCR, or control rabbit IgG antibodies and analyzed by western blotting with astrin, HCR, SKAP, MYCBP and LC8 antibodies. Beta actin antibody was used as negative control; HeLa cell lysates were immunoprecipitated with SKAP, MYCBP and LC8, or control rabbit IgG antibodies and analyzed by western blotting with astrin, HCR, SKAP, MYCBP and LC8 antibodies. Beta actin antibody was used as negative control. Fig. S11. HCR interacts with NuMA. mCherry vector alone or HCR-mCherry was transfected into HeLa cells for immunoprecipitation with mCherry antibody. The precipitates were analyzed by immunoblotting with HCR and NuMA antibodies. Fig. S12. HCR co-localizes with astrin and EDC4.HCR-GFP-transfected HeLa cells were treated with arsenite for 30 min and then co-stained with astrin (red) and EDC4 (cyan) for immunofluorescence detection; scale bars, 10 μm. Fig. S13. HCR co-localizes with EDC4 in mitosis. Negative control, HCR siRNA-treated HeLa cells were co-stained with HCR (red) and EDC4 (green); scale bars, 10 μm. Fig. S14. HCR was closely related to tumorigenesis. The dataset was from TCGA (portal.gdc.cancer.gov) and analyzed by TIMER 2.0 (cistrome.org).Additional file 6. Individual data values. Raw data of Fig. 3A,B,F,G,7C and 7I.Additional file 7. Raw data for Table 1.Additional file 8. Verification reports of HCR-KO cell line.Additional file 9. Verification reports of astrin-KO cell line.Additional file 10. All Original and uncropped blots images used in manuscript. | true | true | true |
|
PMC9590477 | 36304517 | Zhihui Zhang,Jie Wang,Huaqin Kuang,Zhihong Hou,Pingping Gong,Mengyan Bai,Shaodong Zhou,Xiaolei Yao,Shikui Song,Long Yan,Yuefeng Guan | Elimination of an unfavorable allele conferring pod shattering in an elite soybean cultivar by CRISPR/Cas9 | 07-03-2022 | Genome editing,Pod shattering,Soybean,Precision breeding,Marker assisted selection,CRISPR/Cas9 | Pod shattering can lead to devastating yield loss of soybean and has been a negatively selected trait in soybean domestication and breeding. Nevertheless, a significant portion of soybean cultivars are still pod shattering-susceptible, limiting their regional and climatic adaptabilities. Here we performed genetic diagnosis on the shattering-susceptible trait of a national registered cultivar, Huachun6 (HC6), and found that HC6 carries the susceptible genotype of a candidate Pod dehiscence 1 (PDH1) gene, which exists in a significant portion of soybean cultivars. We next performed genome editing on PDH1 gene by clustered regularly interspaced short palindromic repeats (CRISPR)-CRISPR-associated protein 9 (Cas9). In T2 progenies, several transgene-free lines with pdh1 mutations were characterized without affecting major agronomic traits. The pdh1 mutation significantly improved the pod shattering resistance which is associated with aberrant lignin distribution in inner sclerenchyma. Our work demonstrated that precision breeding by genome editing on PDH1 holds great potential for precisely improving pod shattering resistance and adaptability of soybean cultivars. | Elimination of an unfavorable allele conferring pod shattering in an elite soybean cultivar by CRISPR/Cas9
Pod shattering can lead to devastating yield loss of soybean and has been a negatively selected trait in soybean domestication and breeding. Nevertheless, a significant portion of soybean cultivars are still pod shattering-susceptible, limiting their regional and climatic adaptabilities. Here we performed genetic diagnosis on the shattering-susceptible trait of a national registered cultivar, Huachun6 (HC6), and found that HC6 carries the susceptible genotype of a candidate Pod dehiscence 1 (PDH1) gene, which exists in a significant portion of soybean cultivars. We next performed genome editing on PDH1 gene by clustered regularly interspaced short palindromic repeats (CRISPR)-CRISPR-associated protein 9 (Cas9). In T2 progenies, several transgene-free lines with pdh1 mutations were characterized without affecting major agronomic traits. The pdh1 mutation significantly improved the pod shattering resistance which is associated with aberrant lignin distribution in inner sclerenchyma. Our work demonstrated that precision breeding by genome editing on PDH1 holds great potential for precisely improving pod shattering resistance and adaptability of soybean cultivars.
Dear Editor, Conventional breeding has been playing a fundamental role in crop improvement in the past century. Nevertheless, incorporating beneficial genetic variations while excluding unfavorable alleles remains a major challenge, impacted by recombination rates, population size, allelic variations, and effectiveness of phenotypic selection (Lyzenga et al. 2021). As a result, an elite cultivar may take 5–10 years to develop, yet still carry substantial unfavorable allelic variations. Genome editing technologies, particularly clustered regularly interspaced short palindromic repeats (CRISPR)/associated (Cas) nucleases (CRISPR/Cas), can facilitate knock-out, knock-in, and base-editing of target genes, opening up new possibilities to precise improvement of crops (Cai et al. 2020; Chen et al. 2019; Li et al. 2020; Tang et al. 2019; Wang et al. 2020b). The development of CRISPR/Cas-based gene editing has created an avenue for creation of favorable alleles or elimination of maladapted genetic variations in germplasm, before or after the breeding cycle (Lyzenga et al. 2021). Pod shattering has been an unfavorable trait in soybean (Glycine max) breeding, yet is inapparent in humid climate. As a result, a significant portion of soybean cultivars and landraces are still pod shattering-susceptible in areas with higher humidity (Zhang and Singh 2020). The adaptability of shattering-susceptible varieties is severely limited by local climates and is not suitable for introduction to arid regions. For instance, Huachun 6 (HC6) is a national registered cultivar in south China featuring good yield performance and high protein content, yet is shattering-susceptible. The manual harvesting practice and high humidity in south China helped avoid shattering, so yield performance of HC6 is not significantly affected (Fig. 1A). However, in HuangHuaiHai (HHH) region where HC6 can be adapted as summer sowing variety, the low humidity and machine harvest can cause severe yield losses of HC6 (Fig. 1B), limiting its regional adaptability. To diagnose the genetic basis of pod shattering susceptibility in HC6, we performed QTL mapping with a recombinant inbred line (RIL) population of HC6 and pod shattering-resistant JD12. A reproducible major QTL controlling shattering resistance was mapped to chromosome 16, which overlapped with the previously reported qPDH1 QTL (Fig. 1C). The putative PDH1 gene was proposed as Glyma16g25580 (Wm82.a1.v1) encoding a dirigent (DIR) family protein expressed in the inner sclerenchyma of pod walls in shattering susceptible varieties (Funatsuki et al. 2014). We sequenced Glyma16g25580 and found a SNP in JD12 (chr16-29944393, A/T (HC6/JD12); Wm82.a2.v1) leading to a nonsense variant, consistent with a previous report that Glyma16g25580 exists as a truncated gene in shattering-resistant cultivars (Funatsuki et al. 2014). According to the nonsense SNPs A-T, we surveyed the haplotypes of PDH1 gene among resequencing data from 1080 soybean cultivars. The proposed shattering-resistant H-T haplotype is largely fixed in regions with low relative humidity and mechanic harvesting, including 94.50% in northeast China and 85.17% in HHH region (Fig. 1D). In contrast, the shattering-susceptible H-A haplotype is retained in areas with relatively high humidity and/or manual harvesting, including 59.72% of south China, 83.64% of Japan, 81.51% of Korea, and 65.71% of southeast Asia cultivars (Fig. 1D). This result suggested that the presence of Glyma16g25580 gene is highly associated with the relative humidity and harvesting mode shaped pod shattering trait. Genome editing technologies, particularly clustered regularly interspaced short palindromic repeats (CRISPR)/associated (Cas) nucleases (CRISPR/Cas), can facilitate knock-out, knock-in, and base-editing of target genes, creating an avenue for elimination of maladapted genetic variations in germplasm (Cai et al. 2020; Chen et al. 2019; Li et al. 2020; Tang et al. 2019; Wang et al. 2020b). The Glyma16g25580 (designated PDH1 thereafter) open reading frame is mainly responsible for the pod shattering susceptible trait. We then sought to generate mutation of PDH1 by CRISPR/Cas9 in HC6. We designed three sgRNAs, cloned into pGES701 vector individually, and pooled for Agrobacterium-mediated transformation (Fig. 1E–F). Among 23 T0 transgenic plants, 17 lines contained sgRNA, 5 lines contained 2 sgRNAs, and 1 contained all sgRNAs. Hi-TOM (Liu et al. 2019) and sanger sequencing analysis showed that 10 of the 23 T0 transgenic plants carried mutations in at least one target locus. In T1 progenies, we characterized homozygous mutant plants from two lines, HC6pdh1−5 with a 221 bp deletion and HC6pdh1−9 with 1 bp deletion, respectively (Fig. 1G). qRT-PCR showed that the expression of PDH1 gene was diminished by homozygous mutations in both lines (Fig. 1H). In T2 progenies, homozygous mutant plants without transgene were characterized in both lines (data not shown). In a heat dried assay, HC6 pods displayed a high ratio of shattering (Ratio Pod Shattering = 63.75%, n = 8; Fig. 1I–J). In contrast, HC6pdh1−5 and HC6pdh1−9exhibited significant resistance to pod shattering (RPS = 1.25%, n = 8; Fig. 1I–J). The dehisced pod walls of HC6pdh1−5 and HC6pdh1−9 exhibited much lower degrees of torsion than those of HC6, which is consistent with previous finding that qPDH1 locus was associated with curling of pod walls (Funatsuki et al. 2014). We then analyzed the anatomical characteristics of HC6pdh1−5, HC6pdh1−9and HC6 pods. We found that the lignin layer in inner sclerenchyma of HC6 tends to be thicker and looser. In contrast, the lignin layer in HC6pdh1−5 and HC6pdh1−9 appeared to be thinner and compact (Fig. 1L). This result demonstrated that genome editing of the PDH1 gene may affect the pod shattering resistance of HC6 by influencing the deposition of lignin layer in inner sclerenchyma. In 2021 summer, we performed a field trail at Shijiazhuang, Heibei, China. When harvesting was delayed for 2 weeks, HC6 exhibited significant pod shattering that caused substantial yield losses. In contrast, HC6pdh1−5 and HC6pdh1−9 barely showed shattering at the same condition (Fig. 1K). Meanwhile the genome editing of PDH1 did not significantly affect other agronomic traits, including plant height, branch number, pod number per plant, seed number per plant, 100-seed weight, and seed yield per plant (as determined before pod shattering occurs in HC6) (Fig. 1M). Here we showcased the genetic diagnosis and “gene therapy” of the pod shattering trait of soybean by CRISPR/Cas9 genome editing. In comparison with introgression breeding, genome editing approach could rapidly and precisely improve a trait, and is not limited by genetic diversity of breeding populations (Chen et al. 2019; Li et al. 2020; Manghwar et al. 2019). Therefore, genome editing can be integrated as a routine part of a breeding cycle to eliminate unfavorable alleles (such as PDH1) to facilitate the generation of a genetically superior cultivar.
CRISPR/Cas9 mutations of PDH1 in HC6 was performed using the protocol published previously (Bai et al. 2020), and a new CRISPR/Cas9 vector, pGES701 (Fig. 1E), was used for genome editing. The mixed Agrobacterium solution was transformed into the soybean cultivar HC6 via A. tumefaciens-mediated transformation, as described previously (Bai et al. 2020).
To measure the expression of PDH1 gene in WT plants and mutants, real-time qPCR was performed using total RNA extracted from pod wall samples (3 weeks after flowering). Total RNA extraction, cDNA synthesis, and data analysis were performed as previously described (Wang et al. 2020a).
The pod dehiscence percentage of WT and mutations was evaluated by heat treatment: ten fully matured pods of each plant were collected and kept in a circulation drier at 60 °C for 6 h and then counted the number of dehisced pods, respectively. Eight plants per genotype were sampled. Fully matured pods of WT and mutations were examined for pod-wall lignification. Soybean pod was embedded in 7% agarose and cross-sections (80 μm thick) were stained with 10% toluidine blue, and observed under a microscope (Eclipse Ni-U, Nikon, Japan). | true | true | true |
PMC9590522 | 36304421 | Jun Li,Yan Li,Ligeng Ma | Recent advances in CRISPR/Cas9 and applications for wheat functional genomics and breeding | 15-04-2021 | Wheat,CRISPR/Cas9,Genome editing,Functional genomics,Breeding | Common wheat (Triticum aestivum L.) is one of the three major food crops in the world; thus, wheat breeding programs are important for world food security. Characterizing the genes that control important agronomic traits and finding new ways to alter them are necessary to improve wheat breeding. Functional genomics and breeding in polyploid wheat has been greatly accelerated by the advent of several powerful tools, especially CRISPR/Cas9 genome editing technology, which allows multiplex genome engineering. Here, we describe the development of CRISPR/Cas9, which has revolutionized the field of genome editing. In addition, we emphasize technological breakthroughs (e.g., base editing and prime editing) based on CRISPR/Cas9. We also summarize recent applications and advances in the functional annotation and breeding of wheat, and we introduce the production of CRISPR-edited DNA-free wheat. Combined with other achievements, CRISPR and CRISPR-based genome editing will speed progress in wheat biology and promote sustainable agriculture. | Recent advances in CRISPR/Cas9 and applications for wheat functional genomics and breeding
Common wheat (Triticum aestivum L.) is one of the three major food crops in the world; thus, wheat breeding programs are important for world food security. Characterizing the genes that control important agronomic traits and finding new ways to alter them are necessary to improve wheat breeding. Functional genomics and breeding in polyploid wheat has been greatly accelerated by the advent of several powerful tools, especially CRISPR/Cas9 genome editing technology, which allows multiplex genome engineering. Here, we describe the development of CRISPR/Cas9, which has revolutionized the field of genome editing. In addition, we emphasize technological breakthroughs (e.g., base editing and prime editing) based on CRISPR/Cas9. We also summarize recent applications and advances in the functional annotation and breeding of wheat, and we introduce the production of CRISPR-edited DNA-free wheat. Combined with other achievements, CRISPR and CRISPR-based genome editing will speed progress in wheat biology and promote sustainable agriculture.
Common wheat is a keystone crop species. It is grown in many different environments, providing most humans with around 20% of their calories and protein (Uauy et al. 2017); thus, it occupies an important position in food security. As the global population increases, improving the yield of wheat is critical to ensure future availability. Geneticists have exploited natural or artificial wheat variations for breeding. Indeed, conventional breeding approaches have played a major role in increasing grain yields and quality based on broad genetic variations in wheat (Nadolska-Orczyk et al. 2017). However, wheat is an allohexaploid (2n = 6 × = 42, AABBDD); it harbors three closely related subgenomes inherited from three homoeologous ancestors (Petersen et al. 2006). Thus, most wheat genes have three similar but not identical copies, with functional redundancy and complementarity among the A, B, and D genomes. As a result, the probability of the simultaneous mutation of genes in the A, B, and D genomes by natural processes or induced mutagenesis is very low. Therefore, the complex polyploid nature of wheat has hindered the development of functional genomics and breeding, especially compared to other cereals, such as rice and maize. Several genome editing technologies with the ability to change the code of life with high specificity have been developed recently (Li et al. 2019). Advances in genome editing have revolutionized life science, including plant science. The clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9) system offers several advantages, including simplicity, versatility, high efficiency, and the ability to work with multiple targets simultaneously (multiplexing); it has surpassed other genome editing tools, becoming the most widely used gene editing technology in the world (Doudna and Charpentier 2014). Thus far, CRISPR/Cas9 has been used to create various targeted mutations in a broad range of living organisms (Char and Yang 2020; Gürel et al. 2020). Application of the CRISPR/Cas9 system requires the DNA sequences of the target genes. Given the availability of the annotated wheat genome and the elucidation of a growing number of genes controlling important agronomic traits in other plants, it is easy to isolate orthologous genes in wheat based on homology-based cloning. In addition, CRISPR/Cas9 allows researchers to target multiple homoeoalleles simultaneously and it enables the production of targeted mutations in all copies of a gene; thus, the system holds great promise in the characterization of genes endowing important agronomic traits in polyploid wheat. Furthermore, it has been used to modify multiple genes controlling different agronomic traits in wheat. This technology will bring a new dawn to wheat biology and breeding programs. In this review, we briefly outline the utilization of the CRISPR/Cas9 system, with an emphasis on the most important breakthroughs thus far. We also summarize recent applications of genome editing in wheat. Finally, we discuss the future prospects of CRISPR/Cas9 genome editing for wheat improvement.
Genome editing technology generates site-specific double-strand breaks (DSBs) in the targeted genomic sequence using programmable sequence-specific nucleases (SSNs), and then exploits endogenous DSB repair mechanisms to generate a variety of mutations in the target region. Three types of SSNs are used to introduce DSBs at selected sites: zinc-finger nucleases, transcription activator-like effector nucleases, and CRISPR/Cas9 (Kim and Kim 2014; Zhan et al. 2020). Subsequently, the DSBs are mainly repaired via two pathways, nonhomologous end joining (NHEJ) and homologous recombination (HR) (Symington and Gautier 2011). In NHEJ, two broken ends are simply re-ligated, producing insertions and/or deletions (indels) in the target site. When homologous donor sequences are present at the DSBs, HR may be used, and desired gene modifications occur (Chen et al. 2019; Anzalone et al. 2020). Overall, SSN-induced DSBs are repaired more frequently by the NHEJ pathway than by the HR pathway (Carroll 2014; Gao 2021).
CRISPR/Cas systems provide a defense against foreign plasmids or viral DNA elements in bacteria and archaea. They are divided into six types based on the assortment of cas genes and nature of the interference complex (Hille et al. 2018). Three components, mature crRNA, tracrRNA, and Cas9, are responsible for cleaving the invading elements in type II CRISPR/Cas systems. To simplify the system, a dual tracrRNA:crRNA was designed as a single guide RNA (sgRNA) to direct the production of DSBs by Cas9 in vitro (Jinek et al. 2012). Subsequently, an RNA-programmable genome editing tool, CRISPR/Cas9, was developed to create targeted mutations (Cong et al. 2013; Mali et al. 2013; Zhang et al. 2019c). CRISPR/Cas9 contains two major components: a sgRNA, which is responsible for recognizing target DNA, and the Cas9 endonuclease, which is responsible for generating DSB at predesigned target DNA site (Fig. 1A). Cas9 from Streptococcus pyogenes (SpCas9) was the first well-characterized RNA-guided endonuclease. It is a multifunctional protein that contains two nuclease domains: the HNH domain and RuvC-like domain. Each of them cuts one DNA strand, generating blunt-end DSBs; this triggers endogenous DNA repair systems, resulting in targeted mutants. The only prerequisite for applying CRISPR/Cas9 to a given site is the presence of a protospacer-adjacent motif (PAM; NGG for SpCas9) next to the sequence of interest. For different target sites, Cas9 is constant; we can only change the guide sequence in the sgRNA.
Base editing, borrowed from CRISPR, is a precise genome editing approach. It generates targeted point mutations without DSBs, foreign donor templates, or HR (Komor et al. 2016; Gaudelli et al. 2017). Current base editors usually contain a sgRNA, and a catalytically impaired Cas9 nuclease [dead Cas9 (dCas9) or Cas9 nickase (Cas9n)] fused with ssDNA deaminase. The sgRNA guides the modified Cas9-deaminase to the target locus, generating ssDNA R-loop that is exposed and accessible to the deaminase (Anzalone et al. 2020). Based on the different kinds of deaminase, there are two major groups of DNA base editors: cytidine base editors (CBEs) and adenine base editors (ABEs). With CBEs, cytidine deaminase is used to convert cytidine (C) to uridine (U) within the editing window, creating a mismatched base pair with guanine (G) on the opposite strand (Fig. 1B). However, the U intermediate is mutagenic; most organisms have evolved a uracil base excision repair (BER) pathway to excise U from genomic DNA with uracil DNA N-glycosylase. Therefore, uracil glycosylase inhibitor protein (UGI) is used to impede uracil excision, increasing the C-to-T editing efficiency of CBEs (Komor et al. 2016; Nishida et al. 2016). This approach was first used in yeast and human cells and then applied to a variety of plants, including wheat, rice, maize, tomato, and Arabidopsis (Zong et al. 2017; Shimatani et al. 2017; Chen et al. 2017). Theoretically, dCas9/nCas9 fused with an adenosine deaminase would yield an ABE. However, there is no known natural deaminase that deaminates adenine in DNA. By extensive directed evolution and protein engineering of Escherichia coli tRNA adenine deaminase (TadA), researchers produced a deaminase variant (TadA*) that can deaminate adenine (A) in DNA. TadA-TadA* heterodimers were fused with dCas9/nCas9 to generate an ABE (Fig. 1C); adenine is deaminated to inosine (I), treated as G by the polymerase, converting AT to GC base pairs in human cells (Gaudelli et al. 2017). Several groups have shown that ABEs can be applied to plants, including wheat, rice, potato, Arabidopsis, and Brassica napus (Kang et al. 2018; Zong et al. 2018). Many important agronomic traits involve single-nucleotide variants (Zhao et al. 2011; Hu et al. 2015). Therefore, precise editing of a single nucleotide in plants is a desirable and powerful means of accelerating crop improvement. Base editors (CBEs and ABEs) can efficiently mediate all four transition mutations (C → T, A → G, T → C, and G → A) at targeted loci; this will undoubtedly facilitate basic research and breeding in plants.
Prime editing, borrowed from CRISPR, is another precise genome editing method that can generate all 12 types of base substitutions, targeted small insertions, deletions, and combinations of these editing results in the target site. The prime editor mainly consists of a catalytically impaired Cas9 (Cas9n, H840A) fused with an engineered reverse transcriptase and a prime editing guide RNA (pegRNA) (Fig. 1D). The latter contains a primer-binding site at the 3′ end of the sgRNA and a reverse transcriptase template, specifying the target site and also encoding the desired sequence edit (Anzalone et al. 2019). In the prime editing system, the complex binds to the target DNA site and nicks the PAM-containing strand, generating a 3′ end. Then, it hybridizes to the primer-binding site of the pegRNA and primes reverse transcription of the template containing the desired edits. After equilibration between the edited 3′ flap and the unedited 5′ flap, 5′ flap excision, ligation, and DNA repair, DNAs are stably edited in the desired manner (Anzalone et al. 2019). It was first used to correct mutations in human cells and has since been successfully applied to rice plants and wheat protoplasts (Li et al. 2020a; Lin et al. 2020; Xu et al. 2020; Hua et al. 2020). This search-and-replace method is a versatile and precise genome editing tool that does not require DSBs, or donor DNA templates.
When just one homolog mutates, no mutant phenotype might be observed due to masking by other homologs (Borrill et al. 2015). This genetic redundancy and complementarity has hindered the development of wheat biology. Now that CRISPR/Cas9 has been used to target multiple homoeoalleles simultaneously (Table 1), it will accelerate progress in functional genomics and molecular breeding in wheat.
As the world’s population grows, wheat yields will need to substantially increase to ensure global food security. This fact has pushed scientists to investigate and breed innovative wheat varieties. Using CRISPR/Cas9, many negative regulatory genes have been knocked out to improve wheat yields and quality. For example, GASR7 is a gibberellin-regulated gene that controls grain length in rice. Simultaneous targeting of all three TaGASR7 homoeologs significantly elevated the thousand kernel weight, irrespective of the varietal background (Zhang et al. 2016). Similarly, GW2, encoding a RING-type E3 ligase that controls rice grain weight, was knocked out to increase the length and width of wheat grains and, hence, grain yields (Wang et al. 2018b; Zhang et al. 2018). To meet different customers’ needs, grain quality is an important trait that should be improved. For example, gluten proteins, encoded by gliadin genes in wheat, are major factors triggering celiac disease in genetically predisposed individuals. Researchers designed two sgRNAs to target the conserved region of a-gliadin genes; they generated low-gluten wheat, for which the immunoreactivity was reduced by 85% (Sanchez-Leon et al. 2018). Moreover, targeted mutagenesis of TaSBEIIa by CRISPR/Cas9 successfully generated high-amylose wheat with a significantly increased resistant starch content (Li et al. 2020d). Therefore, wheat yields and quality traits can be successfully improved using CRISPR/Cas9.
Heterosis is widely exploited to improve crop productivity and other agronomic traits. However, due to wheat’s strong inbreeding habit, it has been a great challenge to develop male-sterile wheat lines for the production of hybrid seed (Singh et al. 2018). The identification and manipulation of male sterility genes is the first step to generate novel male-sterile wheat mutants. Given the recent molecular identification of male fertility genes in plants, it is possible to clone the homoeoalleles using homology-based cloning, and then create male-sterile hexaploid wheat lines using CRISPR/Cas9. Indeed, great progress has been achieved recently. For instance, NP1 encodes a putative glucose-methanol-choline oxidoreductase that is required for male sterility in rice (Chang et al. 2016). Using an optimized CRISPR/Cas9 system, our group simultaneously disrupted three TaNP1 homoeoalleles in wheat. The resulting Tanp1 triple mutants showed complete male sterility (i.e., produced no pollen) (Li et al. 2020b). Similarly, the targeted knockout of Ms1, which is responsible for pollen exine development and male fertility (Tucker et al. 2017; Wang et al. 2017b), produced complete male sterility in commercial wheat cultivars (Okada et al. 2019). These male sterility mutants could accelerate hybrid breeding of wheat. Very recently, TaCENH3α was edited using CRISPR/Cas9; paternal haploid inducer wheat lines were generated with an induction rate of ~ 7% (Lv et al. 2020). This could be used for additional new breeding technologies and paves the way for reducing the cost of goods in wheat seed production.
Diseases induced by fungi, bacteria, and viruses could reduce wheat yields and quality dramatically. CRISPR/Cas9 has been used to knock out disease-susceptibility genes to generate disease-resistant wheat. In plants, a loss of function of MILDEW-RESISTANCE LOCUS (MLO) confers broad-spectrum resistance to powdery mildew (Gil-Humanes and Voytas 2014). This makes MLO an ideal target for CRISPR/Cas9 to enhance resistance to powdery mildew. Researchers knocked out all six MLO alleles in wheat; they produced a Tamlo triple mutant showing increased resistance to powdery mildew disease (Wang et al. 2014). Similarly, the gene encoding enhanced disease resistance1 (EDR1), a negative factor against powdery mildew defenses, has been simultaneously modified using CRISPR/Cas9, generating wheat with improved powdery mildew resistance (Zhang et al. 2017). Thus, CRISPR/Cas9 is an important means to enhance disease resistance in wheat.
CRISPR/Cas9 is widely used to improve agricultural traits by knocking out unwanted genes or genes conferring undesirable phenotypes. However, this process usually involves transgenic intermediates, which causes regulatory concerns and is not accepted worldwide (Zhang et al. 2020). For public acceptance, gene removal or bypassing foreign elements to edit endogenous genes is a good choice (He and Zhao 2020). Based on the reagents needed for CRISPR-mediated editing, there are two main ways to produce CRISPR-edited DNA-free plants. In the vector-based method, a vector is delivered into wheat callus using Agrobacterium or particle bombardment. It then integrates into the genome and the encoded genome editing elements are expressed, enabling targeted gene knockout. Targeted knockout wheat with foreign DNA is generated in the T0 generation. Ultimately, the foreign DNA can be segregated by selfing and crossing. For example, researchers created a triple-knockout mutant of TaQsd1 via Agrobacterium-delivered CRISPR/Cas9. The mutant was then crossed with wild-type wheat plants, producing transgene-free triple-recessive TaQsd1 mutants that exhibited longer seed dormancy (Abe et al. 2019). Similarly, a marker-free wheat mutant was obtained among the offspring of T0 plants (Wang et al. 2017a). Sometimes, vectors are not integrated into the genome; instead, they may transiently express their encoded genome editing elements to knock out genes. A targeted gene-modified plantlet without foreign DNA is generated in the T0 generation. This approach has been reported in wheat for the first time. Researchers delivered vectors containing CRISPR/Cas9 elements into wheat callus through particle bombardment; the plantlet was subsequently regenerated without antibiotic selection. This transient expression-based CRISPR/Cas9 system produced transgene-free, homozygous mutants (Zhang et al. 2016). In addition, transgene-free wheat carrying nucleotide substitutions have been generated by transiently expressing CBEs or ABEs (Zong et al. 2017; Li et al. 2018). In the non-vector method, Cas9 and sgRNAs are transcribed in vitro and then delivered into immature wheat embryos through particle bombardment. DNA-free genome-edited wheat plants have been generated. Though the editing efficiency was lower, the specificity was higher than with a vector-based system (Zhang et al. 2016). Moreover, nCas9-PBE mRNA and sgRNA were transcribed in vitro and delivered into immature wheat embryos. DNA-free base editing at TaALS-P174 was obtained, endowing wheat with resistance to the herbicide nicosulfuron (Zhang et al. 2019a). In addition, Cas9 can be expressed in vitro and assembled with the sgRNA into a Cas9/sgRNA ribonucleoprotein, which is delivered into immature wheat embryos by particle bombardment. The ribonucleoprotein cleaves the target site immediately and is quickly degraded, generating DNA-free edited wheat (Liang et al. 2017). The final CRISPR-edited DNA-free products are similar to natural and artificial mutants, which are not subject to GMO regulations. We believe that this is the direction of future breeding, and it will play a vital role in realizing sustainable agriculture in the future.
Though large genome and complex polyploid nature have hindered the development of wheat genetic engineering and breeding in the past, several powerful tools are now available to advance wheat biology (Li et al. 2020c). In particular, the development of CRISPR/Cas9 technology has been widely used in wheat genome editing. This technology allows multiplex genome engineering, which has enabled the production of loss-of-function triple wheat mutants; thus, it is a powerful tool for introducing desired traits conferred by a loss-of-function mutation into commercial cultivars via NHEJ. For example, a recessive genic male-sterile (GMS) mutant has several advantages for hybrid wheat production (Li et al. 2020b). As additional genes required for genic male sterility are identified, CRISPR/Cas9-mediated disruption of these genes will enable the rapid production of male-sterile wheat. This represents a promising method for manipulating recessive sterility genes to capture heterosis in wheat. Some valuable alleles are often caused by one or several SNPs or defined insertion/deletions. The introduction of such valuable alleles into commercial cultivars requires 8–10 years by crossing and back-crossing to eliminate unexpected linked traits (Chen et al. 2019). Moreover, breeding is a complicated matter. Sometimes it is not enough to create a good variety by simply modifying one or two genes. For example, five Puccinia graminis f. sp. Tritici (Pgt) resistance genes have been introduced into bread wheat, conferring wheat with broad-spectrum resistance in the field (Luo et al. 2021). Nowadays, CRISPR-mediated precise genome editing is a useful means to achieve these targeted substitutions and replacements by modifying endogenous genes without introducing linkage drag; it can also introduce new alleles (segregating as a single locus) into a predetermined genomic site. Thus, this approach could accelerate the breeding process. Though precise gene modification has been achieved in Arabidopsis, rice, maize, and tomato, it is only feasible in a few laboratories with low efficiency (Li and Xia 2020). To date, except for base editing, precise editing in wheat has not been achieved (Gil-Humanes et al. 2017; Lin et al. 2020). Therefore, precise gene modification in wheat remains a challenge. Given the increased focus of researchers on the mechanism of HR, we believe that precise gene editing via HR will be used for wheat breeding in the near future. The alternative is the dominant repair pathway—NHEJ, which has been exploited to generate gene replacements and gene knock-ins in rice (Li et al. 2016; Dong et al. 2020). It is a promising method to achieve precise genome modification in wheat, and it may facilitate wheat breeding by modifying gene functions or introducing new alleles into a predetermined genomic safe harbor. Transgenerational CRISPR/Cas9 activity has been used to modify multiple target sites in tomato and wheat (Rodriguez-Leal et al. 2017; Wang et al. 2018a). This suggests that valuable, desired phenotypes in elite wheat germplasms, which are recalcitrant to transformation, could be induced by crossing with lines carrying CRISPR/Cas9 elements. In addition, wheat genes have been successfully edited via pollination using CRISPR/Cas9-transgenic maize as a haploid inducer (Kelliher et al. 2020; Budhagatapalli et al. 2020). Such haploid induction-mediated genome editing would not only reduce the genotype dependence on site-specific mutagenesis in wheat, but also provide a path to produce transgene-free gene-edited inbred wheat lines. Collectively, these technologies will accelerate wheat breeding. Some studies have reported that although CRISPR/Cas9 can cleave a target site, sometimes it also cleaves sites with a few mismatches to the target site. This off-target effect is a major concern in gene therapy, but this issue might not be a barrier in plant biotechnology. The putative off-target mutation could be eliminated through back-crossing or crossing with wild-type plants. Moreover, it is advisable to design target sites using web-based tools to reduce off-target mutations by leveraging computation. The fields of genome editing and wheat biology are attracting more and more excellent scientists, and as the number of available CRISPR/Cas platforms increases, additional tools for precisely fine-tuning gene expression will become available in wheat. Combined with other achievements, including the production of high-quality genome sequences and improved transgenic methods, CRISPR and CRISPR-based genome editing will bring functional genomics and rational design-based molecular breeding of polyploid wheat to the forefront of wheat biology. We believe that transgene-free, gene-edited wheat will play a critical role in addressing environmental issues while promoting sustainable agriculture. Significantly, it is not a replacement for traditional breeding; it is just one of the methods advancing wheat breeding programs and accelerating wheat biology. | true | true | true |
PMC9590542 | 36304520 | Jun Li,Xiaoxiao Yu,Chao Zhang,Na Li,Jianjun Zhao | The application of CRISPR/Cas technologies to Brassica crops: current progress and future perspectives | 02-07-2022 | CRISPR,Cas9,Cas12,Precise genome editing,Brassica | Brassica species are a global source of nutrients and edible vegetable oil for humans. However, all commercially important Brassica crops underwent a whole-genome triplication event, hindering the development of functional genomics and breeding programs. Fortunately, clustered regularly interspaced short palindromic repeat (CRISPR)/CRISPR-associated (Cas) technologies, by allowing multiplex and precise genome engineering, have become valuable genome-editing tools and opened up new avenues for biotechnology. Here, we review current progress in the use of CRISPR/Cas technologies with an emphasis on the latest breakthroughs in precise genome editing. We also summarize the application of CRISPR/Cas technologies to Brassica crops for trait improvements. Finally, we discuss the challenges and future directions of these technologies for comprehensive application in Brassica crops. Ongoing advancement in CRISPR/Cas technologies, in combination with other achievements, will play a significant role in the genetic improvement and molecular breeding of Brassica crops. | The application of CRISPR/Cas technologies to Brassica crops: current progress and future perspectives
Brassica species are a global source of nutrients and edible vegetable oil for humans. However, all commercially important Brassica crops underwent a whole-genome triplication event, hindering the development of functional genomics and breeding programs. Fortunately, clustered regularly interspaced short palindromic repeat (CRISPR)/CRISPR-associated (Cas) technologies, by allowing multiplex and precise genome engineering, have become valuable genome-editing tools and opened up new avenues for biotechnology. Here, we review current progress in the use of CRISPR/Cas technologies with an emphasis on the latest breakthroughs in precise genome editing. We also summarize the application of CRISPR/Cas technologies to Brassica crops for trait improvements. Finally, we discuss the challenges and future directions of these technologies for comprehensive application in Brassica crops. Ongoing advancement in CRISPR/Cas technologies, in combination with other achievements, will play a significant role in the genetic improvement and molecular breeding of Brassica crops.
The genus Brassica, belonging to the family Brassicaceae, includes many economically valuable crops that are used as vegetables, oilseeds, and condiments worldwide (Chen et al. 2019a). Six crop species are of particular agricultural importance, and their evolutionary relationships are described by U’s triangle. Three of these species are diploid [Brassica rapa (AA), Brassica nigra (BB), and Brassica oleracea (CC)], while the other three are allotetraploids [Brassica napus (AACC), Brassica juncea (AABB), and Brassica carinata (BBCC)] derived from each pair of the three diploid species. All six Brassica species underwent a recent whole-genome triplication event, resulting in a high number of duplicated genes (Wang et al. 2011). The complex polyploid nature of Brassica has hindered the development of functional genomics and breeding programs. Traditional breeding, molecular marker-assisted selection breeding, and transgenic breeding have been used in Brassica; however, each of them has restrictions or shortcomings (Chen et al. 2019b). Thus, there is significant need to introduce new plant breeding technologies to accelerate germplasm innovation. CRISPR/Cas technology, which allows the editing or modulation of DNA sequences within an endogenous genome, is the most widely used genome-editing technologies (Gao et al. 2021). Considering that single or multiple nucleotide substitutions are crucial for crop improvements (Mao et al. 2019), precise genome-editing platforms based on CRISPR/Cas are highly valuable and have evolved rapidly. These CRISPR/Cas technologies (e.g., base editing, prime editing) have been successfully applied to a broad range of plant species. However, it is predominantly exploited to knock out genes using CRISPR/Cas9 in Brassica species, there is significant room for improvement. Herein, we provide a brief overview of current CRISPR/Cas technologies, including CRISPR/Cas9 and CRISPR/Cas12, and we summarize existing engineered or evolved Cas9 and Cas12a variants with broadened targeting ranges and improved editing specificity. Next, we discuss technical breakthroughs based on CRISPR (e.g., base editing and prime editing), which can carry out precise genome modifications. We also review recent progress in the application of CRISPR/Cas technologies to Brassica species. Finally, current challenges and future perspectives on the use of CRISPR/Cas technologies for Brassica improvement are discussed.
CRISPR/Cas is a prokaryotic adaptive defense system used to fight off invading genetic materials (viruses or plasmids) in bacteria and archaea (Chen et al. 2019b). CRISPR/Cas systems evolve rapidly, leading to extreme structural and functional diversity. Based on the conservation and locus organization of Cas, the systems are divided into two classes: class 1 (types I, III, and IV) and class 2 (types II, V, and VI). With breakthroughs in understanding the defensive processes of CRISPR/Cas systems, CRISPR/Cas9 and CRISPR/Cas12 have been exploited as RNA-programmable genome-editing tools (Fig. 1A–C). Due to their simplicity, high efficiency, versatility, and capacity for multiplexing, CRISPR/Cas technologies have been widely used and have revolutionized all areas of molecular biology (Anzalone et al. 2020; Gao 2021).
Relying on DNA–RNA recognition and binding for targeted DNA breaks, CRISPR/Cas systems are programmed to genome-editing tools (Jinek et al. 2012). These are used to induce site-specific double-strand breaks (DSBs) in the targeted genomic sequence. Once the DSB is made, it triggers two main endogenous DNA repair pathways: non-homologous end joining (NHEJ) and homology-directed repair (HDR) (Fig. 1D). NHEJ is an error-prone pathway. When a DSB is repaired by NHEJ, two broken ends are simply religated, generating uncontrolled insertions/deletions (indels) at the junction of the rejoined chromosome. When pairs of DSBs are created simultaneously, targeted chromosomal rearrangements (e.g., deletions, inversions, and translocations) can be generated between the two breaks (Fig. 1E, F). HDR is a high-fidelity repair process. If a homologous template is provided, HDR may occur, generating precise genome edits, including point mutations, insertions, replacements, and deletions. However, the efficiency of HDR is extremely low in plant cells (Chen et al. 2019b; Mao et al. 2019).
Cas9 from the class 2 type II CRISPR system is an RNA-guided endonuclease. In nature, CRISPR RNA (crRNA) and trans-activating crRNA (tracrRNA) form a two-RNA structure that directs Cas9 to introduce DSB in the target DNA. However, the most widely used CRISPR/Cas9 technology relies on a single guide RNA (sgRNA) engineered from the dual-tracrRNA:crRNA (Jinek et al. 2012). Guided by the target sequence within the sgRNA, Cas9 creates a blunt-ended DSB about 3 bp upstream of the protospacer adjacent motif (PAM, NGG) (Fig. 1A). Since the first report of programmed DNA cleavage by Cas9 from Streptococcus pyogenes (SpCas9) in vitro, Cas9 orthologs have been discovered and tested for genome editing (Gürel et al. 2020). These Cas9 endonucleases differ mainly in their overall size, PAM sequence, guide RNA architecture, and editing efficiency. This has expanded the CRISPR toolbox for genome editing. Due to the high efficiency and versatility, CRISPR/Cas9 has proven to be the best choice for genome editing in numerous species (Chen et al. 2019b; Mao et al. 2019).
Cas12, another RNA-guided endonuclease from the class 2 type V CRISPR system, has also been explored. In nature, many Cas12 endonucleases are guided by a single crRNA while some use an additional tracrRNA (Anzalone et al. 2020). Since the discovery of the mechanism of interference in Acidaminococcus and Lachnospiraceae (Zetsche et al. 2015), CRISPR/Cas12a (formerly named CRISPR/Cpf1) has been adapted for genome editing. Unlike Cas9, Cas12a is guided by a single crRNA, and cleaves targeted DNA distal to T-rich PAM sequences, typically generating DSBs with 4-5-nt staggered overhangs (Fig. 1B). Cas12a orthologs from Acidaminococcus (AsCas12a), Lachnospiraceae (LbCas12a), and Francisella novicida (FnCas12a) have been studied intensively and are commonly used. Although CRISPR/Cas12a is thermosensitive, it is advantageous for multiplex editing (Zetsche et al. 2017); thus, it has become the second leading genome-editing tool. Cas12b has been engineered to cleave both DNA strands (Strecker et al. 2019). Similar to Cas12a, Cas12b prefers T-rich PAMs. Unlike Cas12a, a sgRNA, engineered from a two-RNA structure (crRNA and tracrRNA), directs Cas12b to introduce a DSB with staggered ends (Fig. 1C). Cas12b orthologs have been successfully used for genome editing (Strecker et al. 2019; Ming et al. 2020; Wang et al. 2020a), and CRISPR/Cas12b has become the third most promising RNA-guided endonuclease platform. Recently, other Cas12 nucleases have been explored for use in genome engineering. For example, Cas12e (formerly named CasX) and Cas12j (formerly named CasΦ) are active for eukaryotic genome modification (Liu et al. 2019; Pausch et al. 2020). Of particular interest is Cas12f (formerly named Cas14), which is less than half the size of Cas9/Cas12a. It allows robust gene editing and base editing in mammalian cells (Wu et al. 2021; Xu et al. 2021; Kim et al. 2022); thus, it may be useful in cell engineering and therapeutic applications.
Expanding the target range is key to CRISPR/Cas technology development. Researchers have sought to develop engineered or evolved Cas9 or Cas12a variants with altered or relaxed PAM requirements. Several variants with less restrictive PAM compatibilities have been developed (Anzalone et al. 2020). Of particular interest is the near-PAMless SpCas9 variant, SpRY, which recognizes NRN (R = A or G) and NYN (Y = T or C) PAMs (NRN > NYN), and targets almost all PAMs (Walton et al. 2020). Together, these engineered Cas variants have substantially expanded the range of targets to include those that were previously inaccessible using CRISPR. Improved specificity is another major priority in the development of CRISPR/Cas technologies. Researchers have developed several strategies to enhance the specificity, including exploring sgRNAs with a modified architecture, using dual nickase systems, rationally designing guide RNAs, transiently expressing editing reagents, and delivering editing reagents via preassembled Cas9:sgRNA ribonucleoprotein complexes (RNPs) (Anzalone et al. 2020; Li et al. 2019). Importantly, high-fidelity Cas variants have been rationally engineered or evolved. For example, based on structure-guided protein engineering, eSpCas9(1.1) (Slaymaker et al. 2016) and SpCas9-HF1 (Kleinstiver et al. 2016) have been developed to reduce off-target effects. Furthermore, using the same strategy, a high-fidelity Cas12a variant, enAsCas12a-HF1 (Kleinstiver et al. 2019), has been engineered to improve system specificity. Although these Cas variants have expanded the target range and improved the specificity of CRISPR/Cas, the creation of robust Cas variants through protein engineering remains an important direction for the future advancement of CRISPR technologies. It is also worth to note that design sgRNAs through web tools (e.g., CRISPR-P, CRISPR-GE, CRISPR-PLANT v2, etc.) could reduce or avoid off-targeting (Lei et al. 2014; Minkenberg et al. 2019; Wang et al. 2020b; Xie et al. 2017).
CRISPR-mediated base editing enables direct, irreversible base conversions in a programmable manner, without creating DSBs. Current base editors are fusion proteins composed of catalytically impaired Cas nucleases and single-stranded DNA (ssDNA)-specific deaminases. Guided by guide RNAs, catalytically impaired Cas nucleases localize the ssDNA deaminase to the target sequence, forming a ssDNA R-loop (Anzalone et al. 2020). The nucleotides within the R-loop serve as substrates for the deaminase, and those nucleotide positions define the base editing ‘activity window’ (Komor et al. 2016). To date, two main classes of base editors have been developed: cytosine base editors (CBEs), which convert C–G to T–A base pairs, and adenine base editors (ABEs), which convert A–T to G–C base pairs (Komor et al. 2016; Gaudelli et al. 2017; Gürel et al. 2020). In CBEs, cytidine deaminase is used to directly deaminate cytidine (C) within the ‘activity window’ to uridine (U), resulting in a U–G mismatch. During DNA repair or replication, U is recognized as T, converting C–G to T–A base pairs. Although efficient, targeted C-to-U conversions have been achieved in vitro, the efficiency of base editing in vivo is much lower (Komor et al. 2016). This is probably due to the cellular uridine base excision repair (BER) pathway. Uracil DNA glycosylase (UNG) removes U from DNA in cells and initiates the BER pathway, with reversion of the edited U–G back to a C–G pair. To subvert the BER pathway, uracil DNA glycosylase inhibitor (UGI) was fused to the C-terminus of the CBE architecture. This substantially increased the base editing efficiency. Therefore, CBEs typically include cytidine deaminase, Cas9 nickase, and UGI, and they can catalyze the conversion of C–G to T–A base pairs (Fig. 2A) in various cell types and organisms (Gao 2021). Like cytosine, adenine contains an exocyclic amine that can be deaminated to yield inosine (I), which is read as guanosine (G) by polymerases. In theory, ABEs could be generated by replacing cytidine deaminase with adenine deaminase. However, there are no enzymes known to deaminate adenine in DNA. Therefore, scientists evolved Escherichia coli tRNA adenosine deaminase (TadA) into deoxyadenosine deaminase (TadA*), which can deaminate adenine on ssDNA. Next, TadA* was fused with Cas9 nickase to develop an ABE. To improve the editing efficiency, a wild-type TadA monomer was fused to the N-terminus of the ABE architecture. Among these ABEs, ABE7.10 is recommended for the conversion of A-T to G-C base pairs (Fig. 2B), and its effect has been verified in various cell types and organisms (Gaudelli et al. 2017). Subsequently, monomeric ABE8e has been used in human cells and a variety of species to augment the effectiveness and applicability of adenine base editing (Richter et al. 2020; Li et al. 2021d). A cytosine transversion base editor (CGBE), which converts C to G in human cells and C to A in E. coli (Kurt et al. 2021; Zhao et al. 2021; Chen et al. 2021a), was developed. Structurally, the CGBE is similar to a CBE except that the CGBE contains UNG instead of UGI (Molla et al. 2020). In the CGBE, cytidine deaminase is used to deaminate C directly within the ‘activity window’ to U. Then, it is removed by UNG and creates an apurinic/apyrimidinic (AP) site, initiating the BER pathway. After Cas9 nickase nicks the non-edited strand, DNA repair and replication are activated, realizing the conversion of a C–G to a G–C base pair (Fig. 2C). Recently, the CGBE system has been established in rice, enabling efficient C-to-G editing (Tian et al. 2022). Thus, CGBE expands the base editing toolbox, and helps create new germplasm resources. Dual base editor established by combining CBEs and ABEs has been developed in plants (Li et al. 2020b). It consists of cytidine deaminase, adenosine deaminase, Cas9 nickase, and UGI (Fig. 2D). Guided by a single sgRNA, the cytidine and adenosine deaminases deaminate C to U and A to I within the ‘activity window’, respectively. With DNA repair and replication, this creates C–G to T–A and A–T to G–C base pairs concurrently at the same target site. Dual base editors are valuable tools with broad potential applications, including facilitating the directed evolution of endogenous genes, accelerating the functional annotation of genomes, and aiding the development of therapies for genetic disorders.
Recently, a powerful genome-editing technology named ‘prime editing’ was developed that can install all 12 possible types of base substitutions, small insertions and deletions, and even combinations of these edits (Anzalone et al. 2019). A prime editor consists of Cas9n (H840A) fused to reverse transcriptase (RT) and a prime editing guide RNA (pegRNA), which contains a prime-binding site (PBS) and a RT template at the extended 3' end of the sgRNA. Guided by the pegRNA, Cas9n (H840A) nicks the PAM-containing DNA strand and the PBS hybridizes with the newly liberated 3' end to form a prime-template complex. The RT domain then utilizes the 3' end of the nicked target DNA strand as a primer for reverse transcription. Therefore, the desired edit from the RT template can be permanently incorporated into the target site after excision of the redundant 5' flap and DNA repair of the non-edited strand (Fig. 2E). As a versatile genome-editing tool, prime editing was initially reported to have limited efficiency. Therefore, various strategies have been used to improve its efficiency. For example, PE3 system, which utilizes an additional sgRNA to nick the non-edited strand, increases the editing efficiency ~ 3-fold. However, PE3 produces more undesired indels than PE2. PE3b, in which the added sgRNA targets only the edited sequence, provides a similar editing efficiency and fewer indel byproducts compared with PE3 (Anzalone et al. 2019). Transient co-expression of an engineered DNA mismatch repair (MMR) inhibiting protein has enhanced the editing efficiency. Similarly, in the absence of MMR, PE efficiency is enhanced 2–17-fold (Silva et al. 2022). PEmax systems, which have an optimized editor architecture, can enhance the editing efficiency by an average of 2.8-fold (Chen et al. 2021b). Engineered pegRNAs, with structural RNA motifs or the viral exoribonuclease-resistant RNA motif incorporated into the 3' terminus to enhance the stability, were found to improve the efficiency without increasing off-target editing activity (Nelson et al. 2022; Zhang et al. 2022). pegRNA, split into an sgRNA and a circular RNA RT template, exhibited comparable editing efficiency as that of the pegRNA (Liu et al. 2022). Moreover, various PAM-flexible Cas9 variants have been used in engineered prime editors, not only to increase the number of targetable sites but also to improve the prime editing efficiency (Kweon et al. 2021). Prime editing has also been developed and tested in plants (Li et al. 2020c; Lu et al. 2021); however, the efficiency is limited. Intriguingly, raising the culture temperature to 37 °C increased the editing efficiency an average of 6.3% (Lin et al. 2020). In addition, optimization of the vector component through codon optimization for both Cas9 and M-MLV RT, improving the nuclear localization signal configuration using highly expressed endogenous promoters, and using an enhanced sgRNA scaffold and P2A self-cleaving peptides can improve the editing efficiency at some target sites up to 22-fold (Lu et al. 2021; Xu et al. 2020, 2021). In maize, enhancing pegRNA expression through two pegRNA cassettes or two promoter systems can increase the editing efficiency up to 53.2% (Jiang et al. 2020). Design of the pegRNA is a major determinant of efficiency. For example, designing the PBS sequence with a melting temperature of 30 °C resulted in optimal editing efficiency; a paired-pegRNA strategy encoding the same edits on opposite DNA strands substantially enhanced the efficiency (Lin et al. 2021). Web-based tools (e.g., pegFinder, PlantPegDesigner, PE-Designer, PE-Analyzer, etc.) have been developed to simplify the design of pegRNAs and increase the PE efficiencies (Chow et al. 2021; Hwang et al. 2021; Lin et al. 2021). Recently, engineered Moloney-murine leukemia virus reverse transcriptase (e.g., removing ribonuclease H domain, incorporating a viral nucleocapsid protein, combining both methods) substantially improved prime editing efficiency (Zong et al. 2022). Overall, this revolutionary technology has great potential for plants provided that the editing efficiency is optimized.
CRISPR/Cas technologies can not only mutate a single gene but also mutate multiple genes simultaneously, generating stably inherited knockout mutants in major crops (Mao et al. 2019; Li et al. 2019). The application of CRISPR/Cas technologies to Brassica is rapidly increasing. Initially, marker/reporter genes such as phytoene desaturase were chosen as targets to establish CRISPR/Cas technologies. Since then, diverse endogenous genes have been targeted. Here, we summarize the application of CRISPR/Cas technologies in Brassica (Table 1).
Cabbage (Brassica oleracea var. capitata) is an important leafy vegetable that is grown worldwide. In 2015, CRISPR/Cas9 technology was first applied to cabbage DH1012 by targeting BolC.GA4.a, an ortholog of AtGA4, which is involved in gibberellin biosynthesis. Cas9 driven by a constitutive promoter from Cassava Vein Mosaic Virus, and two sgRNAs targeting the first exon of BolC.GA4.a was constructed in the binary vector. Eighty independent transgenic lines were generated by Agrobacterium-mediated transformation. Through restriction digest/PCR assay, two lines with indels at the target sites were identified out of 20 T0 lines. Additionally, two lines with expected dwarf phenotype were identified through phenotypic screen. Homozygous mutants for BolC.GA4.a showed a dwarf phenotype, and the pod valve margin was affected (Lawrenson et al. 2015). This study demonstrated that CRISPR/Cas9 could induce targeted mutations in cabbage, and that the mutations could be stably transmitted across generations. Due to gene redundancy, simply knocking out one gene may not cause a mutant phenotype. Thus, there is a need to develop tools with the ability to target multiple genes simultaneously. Based on endogenous tRNA processing, researchers developed a CRISPR/Cas9-mediated multiple gene editing system, which can target four sites in a single transformation, and the efficiencies range from 2.8% to 100%. It produced homozygous or biallelic mutations at multiple loci in the T0 generation (Ma et al. 2019a). This system provides an efficient and powerful tool to study gene function and improve traits in cabbage. Since then, a number of studies on CRISPR/Cas9 applications have been published in cabbage. For example, the transcription factor MYB28, a key regulator of aliphatic glucosinolate (A-GSL) biosynthesis, was targeted using CRISPR/Cas9. The myb28 mutant exhibited downregulated A-GSL biosynthesis gene expression and reduced accumulation of methionine-derived glucosinolate (Neequaye et al. 2021). Similarly, researchers obtained a stable cer1 knockout line using CRISPR/Cas9. The genome-edited plant, which had a significantly reduced wax content, had brilliant green leaves (Cao et al. 2021). Together, these studies indicate that CRISPR/Cas9 technology is an important tool for functional genetic studies and the molecular breeding of cabbage.
Chinese cabbage (Brassica rapa spp. pekinensis) is one of the most important vegetables in East Asia. Recently, genome editing was achieved in Chinese cabbage using CRISPR/Cas9 technology. Researchers developed a DNA-free method for the site-directed mutagenesis of endogenous genes using RNPs in protoplasts (Murovec et al. 2018). Meanwhile, BraFLCs, homologs of AtFLC, were targeted using CRISPR/Cas9. Braflc2flc3 double-knockout lines were obtained, and the simultaneous mutations were stably inherited in the T1 generation. The edited plants showed an early-flowering phenotype that was independent of vernalization (Jeong et al. 2019), indicating that CRISPR can be used for molecular breeding. Meanwhile, to confirm the role of BraHINS2 in leaf yellowing, a vector targeting the first exon was constructed and used for transformation. Three T0 lines with targeted mutations were identified from thirteen independent transgenic plants. Mutants homozygous for Brahins2 were obtained in the T1 generation; they were etiolated and senesced at the cotyledon-seedling stage (Su et al. 2021). This study demonstrates that CRISPR/Cas9 can be used to verify the functions of particular genes in Chinese cabbage.
Rapeseed (Brassica napus L.), an allopolyploid crop formed by hybridization between B. oleracea and B. rapa, is one of the most important oil crops worldwide. In 2017, CRISPR/Cas9 was applied to rapeseed. Two BnALC homoeologs were targeted knockout by a CRISPR/Cas9 construct containing only one target sequence, and T1 plants with four alc mutant alleles were obtained. All the mutations were faithfully transmitted to the T2 progeny. Siliques (5–6 cm long) from the alc mutants were more shatter-resistant than same-sized siliques of the cultivar (Braatz et al. 2017). This demonstrates the potential use of CRISPR/Cas9 for the simultaneous modification of genes in a polyploid species. As the applications of CRISPR/Cas9 technology in rapeseed mature, the targeted genes remain diverse but the main focus is on the genetic improvement of commercially important agronomic traits (e.g., yield, nutritional content, and stress resistance). Yield improvement is a main goal of rapeseed breeding that can be increased using CRISPR/Cas9. For example, two sgRNAs were designed to target the two BnaMAX1 homologs. Various mutations were obtained in T0 generation, with the editing efficiency range from 56.30% to 67.38%. The mutations were passed on to the T1 generation. Targeted knockout of two BnaMAX1 genes resulted in high-yield mutants with a significantly decreased plant height and increased branch and silique numbers (i.e., potential rapeseed ideotypes) (Zheng et al. 2020). Similarly, disrupting both copies of BnaCLV3 increased the locule number in siliques, with a significantly higher number of seeds per silique and higher seed weight than in wild-type plants (Yang et al. 2018). Improving the nutrient content of rapeseed is another important goal that can be achieved using CRISPR/Cas9. For example, knocking out all of the copies of BnaFAD2 resulted in an increased oleic acid content in the mutant seeds (Okuzaki et al. 2018; Huang et al. 2020). Likewise, the targeted mutation of BnaTT8 produced tt8 mutants with yellow seeds that contained increased amounts of seed oil and protein, and an altered fatty acid composition (Zhai et al. 2020). Yellow-seeded mutants were also obtained by targeting both copies of BnaTT2; the resulting homozygous mutants exhibited an increased oil content and improved fatty acid composition (Xie et al. 2020). In a final example, CRISPR/Cas9 was used to generate disease-resistant rapeseed plants by targeting WRKY transcription factors. Researchers constructed two vectors with multiple sgRNAs targeting two copies of BnaWRKY11 and four copies of BnaWRKY70, respectively. The resulting wrky70 mutants, but not the wrky11 mutants, exhibited increased resistance to Sclerotinia (Sun et al. 2018). It is worth to note that the phenotype of simultaneous editing multiple duplicated genes is inconsistent with that of editing a single gene sometimes. For example, knocking out all the multiple copies of BnaGTR2 resulted in low seed glucosinolate mutants, however, these cannot be applied for rapeseed breeding because smaller seeds with increased seed oil content were observed (Tan et al. 2022). Very recently, BnaA06.GTR2, a crucial player in seed glucosinolate accumulation, has been targeted knockout. Low seed glucosinolate germplasms were obtained, and no negative effect on yield-related traits were observed (He et al. 2022). Together, CRISPR/Cas9 could not only target a single copy gene but also multi-copy genes, and play important roles for the application in polyploid rapeseed breeding. Besides CRISPR/Cas9, base editors have been successfully applied to rapeseed. For instance, a CBE including rat cytidine deaminase was created that can precisely convert C to T within editing windows from positions 4 to 8 in the protospacer (Wu et al. 2020). Furthermore, a newly developed A3A-PBE system consisting of hAPOBEC3A cytidine deaminase was established. It was more efficient at generating C-to-T conversions within editing windows ranging from C1 to C10 in the protospacer, thus broadening the base-editing window in rapeseed (Cheng et al. 2021) BnaALS was precisely base-edited using a CBE, conferring herbicide resistance to rapeseed (Wu et al. 2020; Cheng et al. 2021). Very recently, CBE has been used to precisely edit ALS in cauliflower; mutants showed strong herbicide resistance (Wang et al. 2022). Another base editor, an ABE, has been applied to rapeseed protoplasts, leading to efficient A-to-G conversions (Kang et al. 2018). These experiments show that CRISPR/Cas technologies can be used to generate valuable germplasm resources for fundamental research on rapeseed and novel variety creation.
CRISPR/Cas technologies show specific, robust, multiplex genome-engineering capabilities. They have been widely used in plant genome editing and played important roles in creating various germplasm resources for crop breeding and biological research (Gao 2021). CRISPR/Cas9 has also been used to create desired mutants in polyploid Brassica species (Table 1). However, there are challenges to its comprehensive application.
The genomes of Brassica species are complex, often with a high ploidy due to their long history of evolution and domestication. Though considerable effort has been made to elucidate the functions of many genes in Brassica, current progress in functional genomics lags far behind that in other crops. Fortunately, with the advent of advanced sequencing technologies and reduced sequencing costs, there has been a dramatic increase in the number of sequenced Brassica genomes (Chen et al. 2019a; Sun et al. 2022). It is easy to identify candidate genes controlling important agronomic traits through homology-based cloning (Karamat et al. 2021). CRISPR/Cas technologies enable multiplex genome editing; therefore, they offer a shortcut to link homologous genes to phenotypes.
Once a gene is selected for editing, genetic transformation is needed to deliver the CRISPR/Cas9 components into plant cells, followed by tissue culture and plant regeneration. Edited plants can then be screened from among the regenerated fertile plants. However, there is no well-established protocol for the genetic transformation of most Brassica crops. Nanotechnology has been proposed as a key driver to address delivery challenges and enhance the utility of plant genetic engineering (Cunningham et al. 2018). For example, nanomaterials enable the delivery of DNA into intact plants, with strong protein expression despite a lack of DNA integration (Demirer et al. 2019). Nanoparticles could potentially be used to deliver editing reagents to Brassica cells. Furthermore, recent studies have reported that overexpressing developmental regulators could improve the efficiency of plant regeneration from tissue culture in various transformation-recalcitrant species and genotypes (Lowe et al. 2016; Nelson-Vasilchik et al. 2018; Debernardi et al. 2020; Kong et al. 2020; Wang et al. 2022). These growth regulators could also be used to generate gene-edited dicots through de novo meristem induction (Maher et al. 2020). Additionally, RUBY has served as a visible and convenient selection marker for transformation in plants. Combining RUBY with CRISPR/Cas9 gene editing cassettes could facilitate to identify gene-edited and transgene-free plants (He et al. 2020). Moreover, a virus-induced genome-editing approach has been developed in wheat, bypassing tissue culture-based transformation (Li et al. 2021c). These strategies promise to expedite progress in genome editing in Brassica species. Agrobacterium-mediated genetic transformation is the most commonly used approach to deliver CRISPR/Cas9 components into plant cells; however, it is restricted to particular genotypes. Recently, the discovery of transgenerational CRISPR/Cas9 activity has facilitated multiple gene editing in plants (Li et al. 2021b). Furthermore, several transformation-recalcitrant crops have been successfully modified using haploid inducer-mediated genome-editing systems (Kelliher et al. 2019; Wang et al. 2019; Budhagatapalli et al. 2020). Importantly, multiple gene homoeologs in B. oleracea and B. napus have been directly modified using a doubled haploid inducer-mediated genome-editing system (Li et al. 2021a). These systems, which enable genome editing in any elite commercial background and can produce transgene-free gene-edited crops when pollinated with (doubled) haploid inducer lines harboring CRISPR/Cas reagents, should be explored for use in transformation-recalcitrant Brassica crops and genotypes. Combined with interspecific hybridization programs, CRISPR/Cas9 will improve the agronomic traits of Brassica crops and accelerate breeding.
Some valuable traits are conferred by single-nucleotide polymorphisms or defined insertions/deletions (Cheng et al. 2016). Thus, harnessing genetic diversity and modifying genomes precisely will be important for crop breeding programs. Until now, apart from the CBEs applied to rapeseed (Ahmar et al. 2022; Wu et al. 2020), precise genome editing (e.g., HDR-based gene targeting, ABEs, and prime editing) had not been achieved in Brassica, and most mutants were obtained via the NHEJ pathway. There is a need for the entry of scientists whose focus is on genome editing into the Brassica field to promote new technologies for precise genome editing. Recently, prime editing and prime editing-based technologies have been used to create point mutations, insertions, fragment deletions, replacements, and inversions (Anzalone et al. 2019, 2021; Choi et al. 2022; Jiang et al. 2022). It will be interesting to apply such versatile and precise technologies to Brassica. Moreover, the efficiency of prime editing systems varies remarkably by target site and cell type (Gao 2021). The DNA repair mechanisms that function in various cell states and cell types, as major determinants of editing efficiency, have not been fully elucidated (Anzalone et al. 2020). Additionally, other issues (e.g., excising redundant 5' flaps, repairing non-edited strands, and preventing byproduct generation) must be addressed. Therefore, effort should be made to optimize prime editing systems; they will facilitate Brassica breeding by modifying gene functions as desired, pyramiding multiple traits, or introducing elite alleles into predetermined safe-harbor loci. Structural variations (SVs) have widespread impacts on the expression of nearby genes; thus, they play an important role in plant evolution and domestication (Alonge et al. 2020). For example, a B. rapa pan-genome was constructed using 18 genomes. Various SVs have been identified and genotyped, revealing the roles of SVs in intraspecific diversification and morphological domestication. Specifically, four SV-related genes are speculated to be involved in leaf-heading domestication (Cai et al. 2021). However, these SVs are not achievable using classical breeding. Chromosome structure engineering (e.g., inversions and translocations) has recently been achieved using CRISPR/Cas technology in plants (Schmidt et al. 2020). Similar systems should be established in Brassica to fix or break genetic linkages, providing huge potential for breeding new varieties with improved traits.
One roadblock to commercializing gene-edited crops is that the process involves genetic modification (Mao et al. 2019; Gao 2021). The current stringent and costly regulation of transgenic genetically modified crops is mainly due to the introduction of foreign DNAs. However, CRISPR/Cas technologies could improve crop traits by altering endogenous genes without transferring transgenes across species boundaries. This may allay fears associated with CRISPR-edited transgene-free plants, reducing the investment in time and money. Due to their low cost and versatility, CRISPR/Cas technologies have been used in various crops, including Brassica. With careful deployment and scientifically informed regulation, DNA-free genome-editing technologies will play important roles in crop improvement programs.
The advent of facile, direct, and precise genome-editing tools using CRISPR/Cas9 has revolutionized plant biology research and crop breeding. Moreover, the expanding knowledge of CRISPR/Cas technologies will continue to be used for innovative applications, which promise to change the pace and course of agricultural research. However, CRISPR/Cas technologies should not be misunderstood as a panacea. Many other achievements are needed as well, including advances in basic genetic research, the development of novel delivery methods, increasing public confidence in the safety of CRISPRed crops, and developing conducive regulatory frameworks. We expect that CRISPR/Cas technologies will be fully applied in Brassica, facilitate the development of functional genomics, and help breed new Brassica varieties with improved agronomic traits. | true | true | true |
PMC9591028 | Letícia Passi Turra,Andressa Romualdo Rodrigues,Fermino Sanches Lizarte Neto,Paulo Cezar Novais,Maria Julia Nunes,Victor Cunha Tirapelli,Fernanda Maris Peria,Vinícius Marques Carneiro,Mucio Luiz de Assis Cirino,Carlos Gilberto Carlotti,Daniela Pretti da Cunha Tirapelli | Expression of microRNAs miR-21 and miR-326 associated with HIF-1α regulation in neurospheres of glioblastoma submitted to ionizing radiation treatment | 19-05-2022 | glioblastoma,ionizing radiation,microRNAs | Background Glioblastoma is an incurable neoplasm. Its hypoxia mechanism associated with cancer stem cells (CSCs) demonstrates hypoxia-inducible factor 1α (HIF-1α) expression regulation, which is directly related to tumor malignancy. The aim of this study was to identify a possible tumor malignancy signature associated with regulation of HIF-1α by microRNAs miR-21 and miR-326 in the subpopulation of tumor stem cells which were irradiated by ion in primary culture of patients diagnosed with glioblastoma. Materials and methods We used cellular cultures from surgery biopsies of ten patients with glioblastoma. MicroRNA expressions were analyzed through real-time polymerase chain reaction (PCR ) and correlated with mortality and recurrence. The ROC curve displayed the cutoff point of the respective microRNAs in relation to the clinical prognosis, separating them by group. Results The miR-21 addressed high level of expression in the irradiated neurosphere group (p = 0.0028). However, miR-21 was not associated with recurrence and mortality. miR-326 can be associated with tumoral recurrence (p = 0.032) in both groups; every 0.5 units of miR-326 increased the chances of recurrence by 1,024 (2.4%). Conclusion The high expression of miR-21 in the irradiated group suggests its role in the regulation of HIF-1α and in the radioresistant neurospheres. miR-326 increased the chances of recurrence in both groups, also demonstrating that positive regulation from miR-326 does not depend on ionizing radiation treatment. | Expression of microRNAs miR-21 and miR-326 associated with HIF-1α regulation in neurospheres of glioblastoma submitted to ionizing radiation treatment
Glioblastoma is an incurable neoplasm. Its hypoxia mechanism associated with cancer stem cells (CSCs) demonstrates hypoxia-inducible factor 1α (HIF-1α) expression regulation, which is directly related to tumor malignancy. The aim of this study was to identify a possible tumor malignancy signature associated with regulation of HIF-1α by microRNAs miR-21 and miR-326 in the subpopulation of tumor stem cells which were irradiated by ion in primary culture of patients diagnosed with glioblastoma.
We used cellular cultures from surgery biopsies of ten patients with glioblastoma. MicroRNA expressions were analyzed through real-time polymerase chain reaction (PCR ) and correlated with mortality and recurrence. The ROC curve displayed the cutoff point of the respective microRNAs in relation to the clinical prognosis, separating them by group.
The miR-21 addressed high level of expression in the irradiated neurosphere group (p = 0.0028). However, miR-21 was not associated with recurrence and mortality. miR-326 can be associated with tumoral recurrence (p = 0.032) in both groups; every 0.5 units of miR-326 increased the chances of recurrence by 1,024 (2.4%).
The high expression of miR-21 in the irradiated group suggests its role in the regulation of HIF-1α and in the radioresistant neurospheres. miR-326 increased the chances of recurrence in both groups, also demonstrating that positive regulation from miR-326 does not depend on ionizing radiation treatment.
Glioblastoma is an incurable neoplasm. Chemoradiotherapy protocols have little effect on survival rates [1]. Therapies based on intrinsic characteristics of tumor stem cells (TSCs) have shown potential results, but the plasticity of the stem state inhypoxia detaches these cells of signals from their niches [2]. Intratumoral hypoxia promotes cancer’s progression from hypoxia-inducible factors (HIFs). Signals from oncogenic transducing pathways regulate the plurality of CSCs and form oncospheres with high mitogenic rates [3]. Cellular bioenergetics is extensively remodeled by intratumoral hypoxia, causing changes in mitochondria oxidation and glycolytic metabolism. When added to an acidic microenvironment, numerous adaptations create favorable conditions for tumor aggression, such as modulations to microRNA expression profiles [3]. Resistance to radiation of glioblastoma is based on HIF-1α regulation. By establishing cooperation with other signaling pathways, the regulation of HIF-1α affects clonogenicity, DNA repair, and cell survival [4]. In severe hypoxic physiological conditions, normal neural progenitors express HIF-1α. This limits therapeutic efforts and reinforces the relevance of studies that analyze the expression of microRNAs, such as oncomiR miR-21 and anti-oncomiR miR-326. Both target HIF-1α regulation in the subpopulation of glioblastoma CSCs submitted to ionizing radiation [3]. miR-21 is substantially considered as oncomiR in glioblastoma. Its high expression is associated with a poor prognosis since, under hypoxic conditions, it can increase the proliferative rate in neural stem cells. However, it has been attested that its overexpression provides pro-differentiation into CSCs, suggesting that miR-21 inhibition could trigger tumor recurrences [3, 5]. The miR-326 and ARRB1 gene are highly expressed in differentiated cells, as they induce cell cycle arrest; other studies indicate that miR-326 induces interleukins 17 (IL-17) by blocking Ets-1, recruits TCD8+ cells promoting cytotoxic activities, in addition, miR-326 directly affects Hh signaling by establishing Smo as a target and reduces Notch pathways 1 and 2; therefore, miR-326 is associated with the capacity for self-renewal and differentiation of CD133+ glioma stem cells and compromises metabolic activity by reducing energy synthesis and cell survival, respectively. Therefore, suggesting that miR-326 may be a marker of glioma aggressiveness [6–9]. Therefore, this study aims to evaluate the expression of microRNAs: miR-21 and miR-326 in a primary culture of patients diagnosed with glioblastoma. We aim to identify a possible tumor malignancy signature associated with regulation of HIF-1α by microRNAs in the subpopulation of tumor stem cells with radiation ionizing.
The present work was developed at the Molecular Biology Laboratory of the Department of Surgery and Anatomy of the Ribeirão Preto Medical School (FMRP-USP).
In this work, we used cell cultures from ten adult patients diagnosed with glioblastoma (glioma grade IV, WHO) at the Neurosurgery Division of the Clinical Hospital, Ribeirão Preto Medical School, University of São Paulo, from August 2014 to April 2015. This project was approved by the Research Ethics Committee of Hospital das Clínicas of Ribeirão Preto (protocol n° 17802/2015) (Tab. 1).
Fresh glioblastoma samples were washed and minced in PBS (1×). This was followed by enzymatic dissociation using collagenase-IV 100 U/mL (Gibco). For the establishment of the neurospheres culture, the cells were suspended in DMEM/F12, with growing factors of EGF and FGF [(20 ng/μL), Life Technologies Corporation, Gaithersburg, MD, USA], and 1% of penicillin. Whereas, for the establishment of the adhered cells culture, these were suspended in DMEM/F12, 10% of bovine fetal serum (Life Technologies Corporation, Gaithersburg, MD, USA) and 1% of penicillin. The cells were put in culture bottles of 75 cm2 (TPP®) and inside an incubator with 5% of CO2 andin 37°C. The culture environment was changed every 48 hours, and the cells were kept in these conditions for about two weeks, until they reached the confluence needed for the execution of the experiments.
After the primary cultures of the ten patients were divided in two initial groups (neurospheres and adhered cells), the infringed experimental procedures were described solely for the neurosphere group: Control group (C): 105 cells were sown and cultivated for 48 hours, and then collected for analysis without treatment [10]. Group treated with ionizing radiation (R): 105 were seeded 24 hours before treatment with ionizing radiation, with a dose rate of 2.0 Gy/min, (60Co source, Unit Gammatron S-80, Siemens, 1.25 MeV, HC-FMRP/USP), for a total dose of 14 Gy. Collection occurred 48 hours after treatment [10].
The cell viability assay was assessed by a dye exclusion method with trypan blue. The cells were briefly re-suspended in 1000 μL of tumor brain stem cell medium, and trypan blue solution (0.4%) was then added to the cells in equal volume. After 15–20 minutes of incubation, we counted the cells using a hemacytometer at room temperature. After counting, the ratio of viable cells to the total number of cells was calculated and recorded.
Total RNA was extracted using Trizol reagent (Applied Biosystems, Foster City, United States) in accordance with the manufacturer’s instructions. Considering the preparation of real-time polymerase chain reaction (PCR), reverse transcription of RNA samples was performed using the High-Capacity cDNA kit (Applied Biosystems). The selected microRNAs were analyzed by the tool MirPath V.3, available on the digital platform Diana Tools (http://snf-515788.vm.okeanos.grnet.gr/). This tool identifies the signals’ pathways that come from the database KEGG (http://www.genome.jp/kegg/pathway.html) which are potentially regulated by the expression of the microRNAs. Thus, in the analysis, it was possible to identify the microRNAs: miR-21-5p and miR-326 which presented as targets the gene HIF-1α. The targets were confirmed in the database of miRTarBase (http://mirtarbase.mbc.nctu.edu.tw/php/download.php). The cDNA was amplified with quantitative real time polymerase chain reaction (q-PCR) using TaqMan Master Mix (Applied Biosystems) for the reaction of microRNAs. The U6 gene was used as an endogenous control (housekeeping) for the reaction of the microRNA. The PCR conditions included pre-heating at 50°C for two minutes, denaturation at 95°C for ten minutes, and 50 cycles of amplification and quantification (15 seconds at 95°C, and one minute at 60°C). All reactions were performed doubly and analyzed with the 7500 Sequence Detection System apparatus (Applied Biosystems). The data were analyzed using ABI-7500 SDS software. Dissociation curves were performed (melting curves) after amplification by RQ-PCR. The samples that showed dissociation curves with different temperatures or more than one point of dissociation in the same sample were not considered and repeated.
Regression coefficients and log transformation application, Poisson regression model, and ROC curve, were used to analyze microRNAs expressions by PCR in real time, relative risks, and predictive value of the studied microRNAs, respectively. All the presented graphs were made with the assistance from software R, version 4.0.0. Analyses were performed with the assistance from SAS 9.2 software. P values smaller than 0.05 were considered to be statically significant.
When analyzing the cell viability of tumor samples with the proposed protocol and isolated radiation by ion, no statistically significant differences were identified between neurospheres from the experimental groups. miR-326 was overexpressed in the group of neurospheres submitted to ionizing radiation treatment when compared to the control group. However, there was no statistically significant difference (p = 0.6309) (Fig. 1A). miR-21 did show statistical difference in the group of neurospheres submitted to ionizing radiation treatment (p = 0.0028) when compared to the control group (Fig. 1B). For every 0.5 units of miR-21, the risk of recurrence increased by 1,004. Thus, there is practically no association between the variables, as the recurrence is not dependent on the occurrence of miR-21 in both groups (Tab. 2A, Fig. 3A, 4A). For every 0.5 units of miR-326, the risk of recurrence increased by 1,024 (2.4%), demonstrating an association between the variables miR-326 and recurrence (p = 0.032) in both groups (Tab. 2A, Fig. 3B, 4B). No associations were verified between the variables miR-21 and miR-326 in relation to mortality in both groups (Tab. 2B, Fig. 3CD, 4CD). The confidence interval shows, with 95% confidence level, the real area under the curve (AUC). Thus, the elevated confidence intervals do not consider any predictions of the studied microRNAs in relation to recurrence and mortality in both experimental groups, despite a near ideal AUC result (Figure 4, 5). The reliability of the estimation also becomes invalid when the AUC is close to the random line, as this indicates a mischaracterization of these microRNAs as a predictive responsive to recurrence and mortality (Fig. 4CD). Consequently, it was not possible to establish cutoff points for sensitivity or specificity (Tab. 5, 6).
miR-21 was overexpressed in the group of neurospheres submitted to ionizing radiation treatment (p = 0.0028) when compared to the control group. It was possible to affirm that miR-21 in normoxic cancer stem cells is one of the interfering factors from radio resistance, which is in agreement with previous research. Hypoxia and the Warburg Effect create an acidic microenvironment that promotes the release of exosomes. In the present study, the neurospheres grew in normoxic conditions, and elevated miR-21 levels in the irradiated group were recorded. These findings suggest radiotherapy eliminated tumor cells and reduced miR-21 levels as a consequence. Thus, ionizing radiation positively corroborated with tumor progression. Other studies that used hypoxic growth conditions have reported that the acid microenvironment triggers the activation of HIF-1α and HIF-2α, stimulating the expression of exosomal miR-21. This substantially promotes the proliferation, migration, and invasion of hepatocellular carcinoma [11]. However, miR-21 was not associated with recurrence and mortality in the present study, as recurrence and mortality occurred independent of miR-21 in both groups. It is suggested that miR-21 does not correlate with putative CSC markers, such as OCT4, SOX2, and CD133. However, the difficulty regarding molecular characterizations of OCT4, SOX2, and CD133 do not exclude the possibility that miR-21 intervenes with the properties of the stem state, suggesting the pro-angiogenic role of miR-21 in glioblastoma. In addition, by immunohistochemistry, negative regulation of PTEN was not recorded in areas that were positive for miR-21, suggesting that miR-21 may not be the principal mechanism that elevated levels of HIF-1α and VEGF [12]. It is extremely rare for glioblastomas to metastasize systematically. However, extracellular vesicles from CSCs are directed to differentiated cells, affecting their phenotype. But the contrary is also true. In glioblastoma, significant concentrations of miR-21 have been detected in the vesicular content, where microglia and monocytes are suppressed. microRNAs derived from extracellular vesicles are not randomly packaged, as they concentrate on only 31–64 target genes. In this scenario, PTEN is silenced too much, and PI3-AKT is activated. Hypoxia is also known to change the vesicular contents in transfer [13]. Studies have shown that miR-21 transferred though extracellular vesicles can deregulate specific mRNAs from wild-type microglia, in vitro and in vivo, as microglial proliferation increases after delivery of miR-21 by negatively regulating BTG2. This decreases the activity of cyclin D1 and reduces cellular proliferation by inhibiting the transition of the G1 phase to the S phase, as demonstrated in fibroblasts. miR-326 expressed positive regulation, but it did not show significant statistical difference when compared to the ionizing radiation treatment group and the control group. However, evidence of an association between miR-326 and recurrence (p = 0.032) was observed in both groups; for every 0.5 units of miR-326, the risk of recurrence increased by 1,024 (2.4%). With this data, it is possible to stress that the expression of miR-326 does not depend on radiotherapy, as ionizing radiation treatment is expected to increase the levels of miR-326. In biological processes, miR-326 is related to the ephrin receptor signaling pathway, axon orientation, and axonogenesis. However, it can also inhibit cell proliferation, migration, and invasion in various types of cancers through the regulation of various factors, such as KRAS, TWIST1, LIM, SH3, ELK1, among other combinatorial systems. Studies have discussed the possibility that the functions and concentrations of miR-326 vary in tumor stages, suggesting that increased levels of this microRNA may correlate with lower survival rates [15]. It is speculated that other molecular agents interact with miR-326 to positively regulate glioblastoma and establish miR-326 as a feedback function [16]. Contrary to the data displayed in this study, the low expression of miR-326 is frequently associated with metastatic development and lower survival rates, as it targets genes related to the activity of disintegrins, metalloproteases, and proteins involving nucleosome dynamics. However, these specific mechanisms have not yet been elucidated [17]. In studies with non-small cell lung cancer, the prognostic function of miR-326 has been correlated with HIF-1α overexpression [15]. Fibroblast growth factor 1 (FGF1) and nerve growth factor (NGF) regulate miR-326 expression. Consequently, this can repair blood vessels by promoting proliferation and invasion of endothelial cells and pericytes. In addition, tumor growth and chemoradiotherapy resistance are stimulated, as MAPK, PI3K, RAS, and JNK mitogenic pathways are signaled. The miR-326high expression, which occurs from decreases of FGF1 and NGF, inhibits malignant behavior. However, abnormal PI3 signaling results in the negative regulation of the miR-326 in glioblastomas [17, 18]. But the association between miR-326 and mortality rates was not verified in the current study. Via the AKT-mTOR pathway, the PINK1 gene is involved in mitochondrial regulation, ATP generation, molecular oxygen consumption, ROS production, and anti-apoptotic and cytoprotective functions. In addition, the low expression of the PINK1 gene decreases the synthesis of glial fibrillary acid protein (GFAP), an astrocyte marker. It is also related to the negative regulation of miR-326. Therefore, the low expression of miR-326 could be related to undifferentiated neural stem cells, as undifferentiated astrocytes demonstrate that PINK1 low expression is not demarcated by GFAP [19]. Proteins of the high mobility groups A1 (HMGA1) and A2 (HMGA2), implicated in tumor invasion and recurrence, are suggested as possible targets for miR-326. Bioinformatic approaches have detected that the oncogenic factor HOTAIR is also a target for miR-326, as its silencing effect significantly reduces fibroblastic growth factor 1 (FGF1) [18]. However, our data demonstrated that the overexpression of miR-326 is a function of the recurrence rate. Such results should encourage investment in new studies with high sample variability to demonstrate the most diverse signals of miR-326, possible target genes, and possible sovereignty over microRNAs that constitute epigenetics and the diverse phenomena of tumorigenesis. Long noncoding RNAs (IncRNAs) can act as sponges to microRNAs, limiting their availability and reducing their regulatory effects. H19 high expression was demonstrated in glioblastomas, and angiogenesis and an increase of invasiveness were confirmed. This contradicts the overexpression of miR-326 in gliomas, and it results in decreased cell proliferation and an apoptotic increase [16]. The area under the curve (AUC) in C groups, in both microRNAs, is considerable. However, the small sample size does not allow us to stress adequate cutoff points for recurrence and death. However, tumor heterogeneity increases population variability. Therefore, a large CI would be an inherent aspect of malignancy in glioblastoma. Solely from this study, it is not possible to define microRNAs miR-21 and miR-326 as prognostic markers for stage assessment in radiotherapy. The sample number must be increased in future studies so that cutoff points can reach maximum diagnostic and prognostic accuracy. The experimental design should also be applied to liquid biopsies from plasma and cerebrospinal fluid. In the present study, cell growth occurred in normoxic conditions. However, studies have suggested that significant tumor growth occurs after radiotherapy in regions with high incidences of intermittent vascular stasis. Therefore, cyclically hypoxic cells may be more radioresistant for modulating angiogenic processes, which also characterizes radioresistance to the tumor endothelium20. Future experiments should be performed in hypoxic conditions to allow for findings compatible with the pO2 of solid tumors in vivo. Our work confirms the role of miR-21 in the regulation of the subpopulation of tumor stem cells. Our results also reinforce that the concentration and function of microRNAs, such as miR-326, have a unique signature in different tumor stages, relapses, and responses to treatments. Most importantly, future research should investigate the diverse interferences that overlap the regulatory effects of tumor-suppressing microRNAs. Investment is needed in research that constitutes a therapeutic set based on maneuvers related to tumor oxygenation, degradation of HIF-1α e HIF-2α, inhibition of residual HIF-1α, and associating these factors with immunotherapy to silence glioblastoma malignancy.
In summary, miR-21, under normoxic conditions, was overexpressed in the group of neurospheres submitted to ionizing radiation treatment. This suggests that its role in the regulation of HIF-1α is related to radioresistant neurospheres. miR-326, under normoxic conditions, was associated with an increased risk of recurrence in both groups, also demonstrating that this positive regulation does not depend on ionizing radiation treatment. | true | true | true |
|
PMC9591066 | Marie-Sophie Friedl,Lara Djakovic,Michael Kluge,Thomas Hennig,Adam W. Whisnant,Simone Backes,Lars Dölken,Caroline C. Friedel | HSV-1 and influenza infection induce linear and circular splicing of the long NEAT1 isoform | 24-10-2022 | The herpes simplex virus 1 (HSV-1) virion host shut-off (vhs) protein cleaves both cellular and viral mRNAs by a translation-initiation-dependent mechanism, which should spare circular RNAs (circRNAs). Here, we show that vhs-mediated degradation of linear mRNAs leads to an enrichment of circRNAs relative to linear mRNAs during HSV-1 infection. This was also observed in influenza A virus (IAV) infection, likely due to degradation of linear host mRNAs mediated by the IAV PA-X protein and cap-snatching RNA-dependent RNA polymerase. For most circRNAs, enrichment was not due to increased circRNA synthesis but due to a general loss of linear RNAs. In contrast, biogenesis of a circRNA originating from the long isoform (NEAT1_2) of the nuclear paraspeckle assembly transcript 1 (NEAT1) was induced both in HSV-1 infection–in a vhs-independent manner–and in IAV infection. This was associated with induction of novel linear splicing of NEAT1_2 both within and downstream of the circRNA. NEAT1_2 forms a scaffold for paraspeckles, nuclear bodies located in the interchromatin space, must likely remain unspliced for paraspeckle assembly and is up-regulated in HSV-1 and IAV infection. We show that NEAT1_2 splicing and up-regulation can be induced by ectopic co-expression of the HSV-1 immediate-early proteins ICP22 and ICP27, potentially linking increased expression and splicing of NEAT1_2. To identify other conditions with NEAT1_2 splicing, we performed a large-scale screen of published RNA-seq data. This uncovered both induction of NEAT1_2 splicing and poly(A) read-through similar to HSV-1 and IAV infection in cancer cells upon inhibition or knockdown of CDK7 or the MED1 subunit of the Mediator complex phosphorylated by CDK7. In summary, our study reveals induction of novel circular and linear NEAT1_2 splicing isoforms as a common characteristic of HSV-1 and IAV infection and highlights a potential role of CDK7 in HSV-1 or IAV infection. | HSV-1 and influenza infection induce linear and circular splicing of the long NEAT1 isoform
The herpes simplex virus 1 (HSV-1) virion host shut-off (vhs) protein cleaves both cellular and viral mRNAs by a translation-initiation-dependent mechanism, which should spare circular RNAs (circRNAs). Here, we show that vhs-mediated degradation of linear mRNAs leads to an enrichment of circRNAs relative to linear mRNAs during HSV-1 infection. This was also observed in influenza A virus (IAV) infection, likely due to degradation of linear host mRNAs mediated by the IAV PA-X protein and cap-snatching RNA-dependent RNA polymerase. For most circRNAs, enrichment was not due to increased circRNA synthesis but due to a general loss of linear RNAs. In contrast, biogenesis of a circRNA originating from the long isoform (NEAT1_2) of the nuclear paraspeckle assembly transcript 1 (NEAT1) was induced both in HSV-1 infection–in a vhs-independent manner–and in IAV infection. This was associated with induction of novel linear splicing of NEAT1_2 both within and downstream of the circRNA. NEAT1_2 forms a scaffold for paraspeckles, nuclear bodies located in the interchromatin space, must likely remain unspliced for paraspeckle assembly and is up-regulated in HSV-1 and IAV infection. We show that NEAT1_2 splicing and up-regulation can be induced by ectopic co-expression of the HSV-1 immediate-early proteins ICP22 and ICP27, potentially linking increased expression and splicing of NEAT1_2. To identify other conditions with NEAT1_2 splicing, we performed a large-scale screen of published RNA-seq data. This uncovered both induction of NEAT1_2 splicing and poly(A) read-through similar to HSV-1 and IAV infection in cancer cells upon inhibition or knockdown of CDK7 or the MED1 subunit of the Mediator complex phosphorylated by CDK7. In summary, our study reveals induction of novel circular and linear NEAT1_2 splicing isoforms as a common characteristic of HSV-1 and IAV infection and highlights a potential role of CDK7 in HSV-1 or IAV infection.
Herpes simplex virus 1 (HSV-1) is one of nine herpesviruses known to infect humans [1, 2]. It is most commonly known for causing cold sores but can also result in life-threatening diseases. A key role in HSV-1 lytic infections is played by the virion host shut-off (vhs) protein, which is delivered to the infected cell by the incoming virus particles and rapidly starts cleaving both cellular and viral mRNAs [3]. Vhs target recognition occurs during translation initiation and requires binding of vhs to components of the cap-binding complex eIF4F [4–8]. Vhs can also cleave circular RNAs (circRNAs), but only if they contain an internal ribosome entry site (IRES) [9]. CircRNAs form covalent RNA circles and are naturally generated during the splicing process either by “back-splicing” of exons out of their linear order (resulting in non-colinear 3′–5′ junctions) or stabilization of intron lariats with 2′–5′ junctions joining the 5′ and branchpoint nucleotides (Fig 1A) [10]. Back-splicing requires canonical splice sites and spliceosome assembly [11]. Previously considered to occur only rarely [12], recent large-scale RNA sequencing (RNA-seq) studies revealed the existence of thousands of circRNAs in eukaryotic cells [13, 14]. Due to the absence of a 5’ cap or poly(A) tail and their resistance to exonucleases, circRNAs are more stable than linear RNAs [10]. While a function as miRNA sponge has been reported for a small number of circRNAs [15, 16], this is now nevertheless thought to be rare [14]. Other functions reported for individual circRNAs involve regulation of transcription, alternative splicing and translation and some may even serve as templates for protein synthesis [14]. However, for most circRNAs their function remains elusive. Recently, Shi et al. reported on dysregulation of circRNAs during HSV-1 infection, with 188 circRNAs being significantly up-regulated at 48 h post infection (p.i.) [19]. Their differential analysis was performed on FPKM values (= fragments per kilobase of transcript per million fragments mapped) calculated for circRNAs from reads crossing the circular, i.e., back-splicing, junction. For FPKM normalization, circRNA read counts were divided by the total number of reads aligning to the host genome, most of which map to linear transcripts. Thus, this analysis cannot distinguish between a true up-regulation of circRNA biogenesis and an enrichment of circRNAs in case of a selective loss of linear RNAs. Considering that vhs cleaves translated linear transcripts but not circRNAs without an IRES, we hypothesized that the seeming up-regulation of many circRNAs in HSV-1 infection largely represented the escape of circRNAs from vhs-mediated RNA decay. This would result in an enrichment of circRNAs similar to, but likely less pronounced than RNase R treatment commonly used to enrich circRNAs [20]. We previously performed 4-thiouridine-(4sU)-tagging followed by sequencing (4sU-seq) to characterize de novo transcription in hourly intervals during the first 8 h of lytic wild-type (WT) HSV-1 strain 17 infection of primary human foreskin fibroblasts (HFFs) and combined this with total RNA-seq in two-hourly steps [21, 22] (Fig 1B, n = 2 replicates). More recently, we applied the same experimental set-up to Δvhs infection to study vhs-dependent gene regulation [18] (Fig 1B, n = 2 replicates). In this study, we explore these data to show that vhs-mediated degradation of linear RNAs in HSV-1 infection leads to a general enrichment of circRNAs relative to linear circRNAs. This not only confounds analysis of differential circRNA expression but also standard differential gene expression and exon usage analyses. Our analysis also revealed actual up-regulation of both circular and novel linear splicing of the long non-coding (linc)RNA nuclear paraspeckle assembly transcript 1 (NEAT1) during HSV-1 infection. In contrast to other circRNAs, this induction was independent of vhs and resulted from de novo synthesis of both the circRNA and linear splicing isoforms in HSV-1 infection. The nuclear lincRNA NEAT1 is one of the most highly expressed lincRNAs and has two isoforms with identical 5’ but distinct 3’ ends [23] (Fig 1C): A short transcript (NEAT1_1, also denoted as MENε, ~3.7nt), stabilized by a poly(A) tail and expressed in a wide range of cells [24], and a long transcript (NEAT1_2, also denoted as MENβ, ~22.7 nt), stabilized by a triple helical structure [25] but with more limited expression in particular cell types [24]. The HSV-1-induced circRNA is generated from the 3’ region unique to the NEAT1_2 transcript (Fig 1C). NEAT1_2 is essential for the structure of paraspeckles [26], nuclear bodies located in the interchromatin space. NEAT1_2 forms a scaffold for paraspeckles by being bound by a number of proteins with functions in transcription and/or RNA processing and splicing (reviewed in [27, 28]). Paraspeckles impact gene expression by nuclear retention of adenosine-to-inosine-edited mRNAs [28]. In viral infection, NEAT1_2 plays a role in interleukin 8 (IL-8) induction by sequestering the IL-8 repressor splicing factor proline/glutamine-rich (SFPQ) to paraspeckles [29]. NEAT1 is up-regulated in HIV-1, dengue, HSV-1, Hantavirus, Hepatitis D and influenza infections [28]. Expression levels of NEAT1_2 have been shown to be positively correlated with presence of paraspeckles [30] and HSV-1 infection induces paraspeckle formation, likely through up-regulation of NEAT1 [31]. The HSV-1 immediate-early proteins ICP27 and ICP22 are both known to interact with the host transcription and RNA processing machinery. ICP27 regulates splicing and polyadenylation by interacting with splicing factors and the mRNA 3’ processing factor CPSF [32–35]. We recently demonstrated that ICP27 is sufficient but not necessary for disruption of transcription termination in HSV-1 infection [34]. ICP22 represses RNA Polymerase II (Pol II) transcription elongation by interacting with elongation factors, such as the positive transcription elongation factor b (P-TEFb) and the FACT complex, and inhibits phosphorylation of the Pol II carboxyterminal domain (CTD) at Ser2 residues [36]. Ser2 phosphorylation plays a key role in recruiting splicing and termination factors to Pol II [37]. We thus hypothesized that ICP27 and/or ICP22 may play a role in induction of NEAT1_2 splicing. Indeed, we could show that combined ectopic expression of ICP27 and ICP22 is sufficient to induce both circular and linear splicing of NEAT1_2, however neither protein was required. HSV-1 and influenza A (IAV) infection as well as heat stress lead to NEAT1_2 up-regulation [29, 38] and read-through transcription beyond poly(A) sites for tens-of-thousands of nucleotides for many but not all cellular genes [21, 22, 39–41]. IAV infection, but not heat stress, also induces NEAT1_2 circular and linear splicing. Moreover, selective inhibition and knockdown of cyclin-dependent kinase 7 (CDK7) as well as knockdown of its phosphorylation target the MED1 subunit of the Mediator complex [42] induce both NEAT1_2 splicing and read-through transcription in cancer cell lines. This highlights a possible link between read-through transcription and NEAT1_2 splicing and a potential role of CDK7 in HSV-1 or IAV infection.
CircRNA detection from RNA-seq reads is based on identifying junction reads connecting a donor splice site of a downstream exon to the acceptor splice site of an upstream exon. Several algorithms for circRNA de novo detection are available, but due to high false positive rates combination of at least two algorithms is recommended [43]. We thus employed a pipeline combining circRNA_finder [44] and CIRI2 [45] for circRNA de novo detection (see methods and Fig 2A) and analyzed only circRNAs that were independently discovered by both algorithms. We applied this pipeline on our total RNA-seq and 4sU-seq time-courses for the first 8 h of lytic wild-type (WT) and Δvhs HSV-1 infection (Fig 1B). This identified a total of 16,463 circRNAs in the WT time-course and 7,306 in the Δvhs infection time-course, with 1,647 to 6,887 circRNAs detected per sample in total RNA (S1A Fig). In contrast, only a small number of circRNAs were detected in 4sU-RNA (119 and 194 in at least one sample of WT and Δvhs 4sU-seq, respectively, S1B Fig). This is partly consistent with the study by Zhang et al. who also obtained smaller numbers of circRNAs (255, 820, 876 circRNAs in PA1, hESC H9 and H9 differentiated FB Neurons cells, respectively) with 1 h of 4sU-tagging using 4sUDRB-seq (DRB treatment followed by 4sU-seq after DRB removal) compared to total RNA-seq (6,740, 4,523, and 11,185 circRNAs, respectively) [46]. While this highlights a high variability between experiments and/or cell types in the number of circRNAs detected, the higher number of circRNAs detected by Zhang et al. can also partly explained by the use of only one circRNA detection algorithm (CIRCexplorer [47]). This is less restrictive than requiring independent identification of circRNAs by two algorithms as done in our study, but likely has a higher false positive rate. Moreover, circRNA_finder and CIRI2 require presence of canonical GT-AG splicing signals, while CIRCexplorer also identifies potential circRNAs with other less common splicing signals. Most circRNAs in our study (>97%) were only detected in total RNA and almost all of the identified circRNAs (82.6% and 92.7% for the WT and Δvhs infection time-courses, respectively) were already annotated in circBase, a database of circRNAs identified in large-scale RNA-seq studies [48]. In total RNA, a substantially larger number of circRNAs was identified for WT infection than for mock and Δvhs infection (S1A Fig). Since the number of reads mapping to the host genome in total RNA was comparable or higher in Δvhs infection than in WT infection (S1C Fig), adjusting the read count threshold for a circRNA to be detected to sequencing depth further increased the difference in detected circRNAs between WT and Δvhs infection (Fig 2A). Thus, the lower number of circRNAs identified in Δvhs infection is not due to lower sequencing depth. Moreover, normalized to sequencing depth, a smaller number of circRNAs was detected in uninfected cells of the Δvhs infection time-course compared to the WT infection time-course. In part, this could be due to shorter read length (76 nt in the Δvhs infection time-course vs. 101 nt in the WT infection time-course, both paired-end sequencing), the effect of which will be further investigated below. However, other experimental differences between the two time-courses, which were performed at different times, likely contribute. To account for these experimental differences, we always compared samples of WT infection against mock infection of the WT infection time-course and samples of Δvhs infection against mock infection of the Δvhs infection time-course. To evaluate enrichment of individual circRNAs relative to linear transcripts in HSV-1 infection, circRNA read counts were normalized to the total number of reads mapping to linear exon-exon junctions of all human protein-coding genes (not only the “parent” genes from which circRNAs originate) and then averaged between replicates. The reason we did not normalize only against reads from “parent” genes is that reads that map linearly to splicing junctions within circRNA regions could originate from either linear or circular transcripts. Excluding those reads would result in low numbers of reads for normalization, leading to very noisy estimates of enrichment, in particular for HSV-1 infection where linear transcripts are expected to be lost. Including those reads, however, would lead to under-estimation of circRNA enrichment. Since only a relatively small fraction of genes harbour circRNAs, use of junction reads from all genes circumvents this issue. Moreover, there are circRNAs without corresponding linear transcripts, such as CDR1as, one of the few circRNAs with a well-described function [16]. This circRNA would otherwise have to be excluded from our analysis. Notably, the fraction of linearly spliced host reads decreased considerably during WT infection (S1D Fig). This is largely a consequence of HSV-1-induced read-through transcription. Read-through transcripts are rarely spliced downstream of the poly(A) site and are retained in the nucleus [21, 22], thus escaping vhs-mediated decay. In contrast, in vhs infection, the fraction of spliced reads did not decrease (S1E Fig) due to reduced read-through transcription and absence of vhs-mediated decay of cytosolic mRNAs. Thus, normalization by linear splice junction read counts better quantifies the impact of vhs-mediated decay on linear transcripts than normalization by the total number of reads mapped to the host genome as in WT infection a large fraction of host genome reads originate from read-through transcripts that escape vhs-mediated decay. Normalized circRNA counts increased substantially with increasing duration of WT infection, indicating increasing enrichment of circRNAs relative to linear transcripts (Fig 2C, S2A–S2C Fig, red line = linear regression estimate over all circRNAs). By 8 h p.i. WT infection, circRNAs were enriched ~6.5-fold in total RNA compared to uninfected cells. In contrast, Δvhs infection exhibited only a very modest ~1.3-fold enrichment in circRNAs by 8 h p.i. (Fig 2D, S2D–S2F Fig). This can likely be attributed to the global loss of host transcriptional activity in HSV-1 infection [51], which affects circRNAs less than linear transcripts as circRNAs are much more stable. Notably, if we instead normalize by the total numbers of reads mapped to the host (as done in the study by Shi et al. [19]), we also observe enrichment of circRNAs in WT infection, albeit to a lower degree (S3 Fig). As noted above, one factor that might negatively impact circRNA detection in the Δvhs infection time-course is the shorter read length. CIRI2 and circRNA_finder require that at least one read overlaps by at least 19 and 20 nt, respectively, on both sides of identified circular junctions. Thus, longer reads could potentially lead to identification of more circRNAs. To address this issue, we pursued two alternative approaches. First, we trimmed reads from the WT infection time-course to 76 nt and applied our de novo circRNA detection method to these trimmed reads. Second, we implemented an alternative circRNA detection method based on aligning reads against circRNA junction sequences constructed from annotated circRNAs, which allowed to reduce or increase the required overlap on either side of the junction (see methods and S4 Fig). In this case, false positive rates are reduced by the restriction to annotated circRNAs. Both with trimmed reads and the alignment-based pipeline with a required overlap of only 10 nt, we observed approximately the same circRNA enrichment in WT infection as with the de novo detection pipeline on full reads (S5 and S6 Figs). Moreover, since circRNA enrichment is consistent between highly and lowly expressed circRNAs as shown by the low deviation from the linear regression line, exclusion of lowly expressed circRNAs by adapting circRNA detection thresholds to sequencing depth would not alter conclusions. Unless explicitly noted otherwise, all results reported in the following use the alignment-based pipeline requiring an overlap of at least 10 nt on either side of the junction. We also analyzed data from a separate RNA-seq experiment in which we obtained total RNA for mock, WT, Δvhs and ΔICP27 infection in the same experiment as well as subcellular fractions (chromatin-associated, nucleoplasmic and cytosolic RNA) (all 8 h p.i., n = 2) [18, 22, 52]. This showed enrichment of circRNAs in total RNA of WT and ΔICP27 infection, but not Δvhs infection (S7 Fig), Notably, although linear regression analysis over all circRNAs (red line) estimated only a 1.5- to 2-fold enrichment in WT infection, some of the most highly expressed circRNAs were more strongly enriched. To confirm that enrichment of circRNAs in HSV-1 infection was not due to increased abundance of circRNAs, we also analyzed recently published total RNA-seq data for mock and HSV-1 strain KOS (WT KOS) infection (n = 1 replicate), where ERCC spike-ins were added to allow absolute quantification of expression levels [53]. This confirmed the strong enrichment of circRNAs relative to linear RNAs in a second HSV-1 strain (~10-fold, S8A Fig). However, normalization to ERCC spike-in read counts showed that absolute levels of most circRNAs in total RNA remained largely unchanged during HSV-1 infection (S8B and S8C Fig). We conclude that vhs-mediated degradation of linear mRNAs that spares circRNAs leads to an enrichment of circRNAs relative to linear mRNAs during HSV-1 infection. As circRNA reads that do not cross circular junctions can be aligned by standard RNA-seq mapping programs, they are counted towards circRNA “parent” genes and exons in standard differential gene expression (DGE) and differential exon usage (DEU) analyses. Consequently, genes encoding for circRNAs tend to have significantly higher fold-changes in DGE analysis of total RNA in WT infection from 6 h p.i. (one-sided Wilcoxon rank sum test, p-value <0.0001, S9A Fig). This was not observed in Δvhs infection. Similarly, DEU analysis with DEXSeq showed differential exon usage in WT infection for many exons contained in circRNAs (multiple testing adjusted p-value ≤0.005). By 8 h p.i. WT infection, ~17% of circRNAs showed differential exon usage for at least one exon within the genomic range of the circRNA (Fig 2E), with increased exon usage for almost all of these circRNAs (Fig 2F). This effect was most pronounced for highly expressed circRNAs (S9B Fig). Consistent with the small degree of circRNA enrichment late in Δvhs infection, 3.6% of circRNAs also showed differential exon usage by 8 h p.i. Δvhs infection (S9C and S9D Fig). In summary, our results demonstrate that enrichment of circRNAs in WT infection biases not only differential circRNA analyses, but also DGE and DEU analyses of total RNA and thus needs to be taken into account for the latter.
In contrast to other circRNAs, one circRNA (circBase ID: hsa_circ_0003812) was actually induced in HSV-1 infection even in absence of vhs. hsa_circ_0003812 is encoded in the unique part of the long NEAT1_2 transcript (Fig 1C), was essentially absent in mock-infected cells (0–2 reads in 4sU-RNA, total RNA and all subcellular fractions, ≤0.22 normalized circRNA count) and highly abundant in HSV-1 infection (Fig 3A, S10 Fig). ERCC spike-in normalization in mock and WT KOS infection confirmed that abundance of this NEAT1_2 circRNA indeed increased in absolute levels (S8B and S8C Fig), indicating high levels of de novo synthesis of this circRNA in HSV-1 infection. Consistent with this, induction of the NEAT1_2 circRNA during HSV-1 infection was also confirmed in samples of very recently transcribed RNA, i.e., 4sU-, nucleoplasmic and chromatin-associated RNA. This is particularly noteworthy considering the small number of circRNAs we recovered in these samples (S1B, S1C, S11 Figs). From 6 h p.i., the NEAT1_2 circRNA was among the most highly expressed circRNAs in total, 4sU-, chromatin-associated and nucleoplasmic RNA in WT, ΔICP27 and Δvhs infection (S10 Fig). Consistent with the nuclear localization of NEAT1, only few circular NEAT1_2 reads were found in cytosolic RNA. Other NEAT1_2 circRNAs were also detected, but only with much fewer reads (Fig 3A). In summary, our data indicate that biogenesis of the hsa_circ_0003812 NEAT1_2 circRNA is induced during HSV-1 infection. The hsa_circ_0003812 NEAT1_2 circRNA was previously found to be enriched in RNA-seq data of human fibroblasts after treatment with the 3’->5’ RNA exonuclease RNase R [54], confirming that it is a circRNA. Notably, however, NEAT1_2, is partially resistant to RNase R due to the stabilizing triple helical structure at its 3’end, similar to MALAT1 [20]. Thus, the linear NEAT1_2 transcript is not strongly depleted by RNase R treatment even with an improved protocol using A-tailing prior to RNase R treatment and a Li+ buffer (S12A Fig) [20]. Nevertheless, we identified a study of circRNAs in Akata cells, in which the linear NEAT1_2 transcript was fully depleted by RNase R treatment in total and cytoplasmic RNA, yet partially resistant in nucleoplasmic RNA, while the hsa_circ_0003812 circRNA was always retained (S12B Fig) [55]. Interestingly, we observed significant enrichment of the circRNA region in our total RNA-seq time-course of WT infection from 6 h p.i. compared to both up- and downstream regions (Fig 3B, S13A Fig, 1.5- to 1.76-fold increase in the DEXSeq analysis, multiple testing adjusted p<0.005). This effect is likely not mediated by vhs since NEAT1_2 is located in the nucleus while vhs is active in the cytosol. Furthermore, some enrichment of the circRNA region was also observed in Δvhs infection by 8 h p.i., consistent with an attenuated progression of Δvhs infection (S13B Fig). Both NEAT1_1 and NEAT1_2 were found to be highly unstable in mouse, with half-lives of ~30 and ~60 min, respectively [56]. Thus, enrichment of the NEAT1_2 circRNA in HSV-1 infection is likely due to its high stability compared to the linear transcript combined with high de novo synthesis of this circRNA. The following evidence further confirms that hsa_circ_0003812 is a circRNA. First, it was identified by three distinct algorithms for circRNA detection that employ different approaches: CIRI2, circRNA_finder and our aligment-based approach. Combination of two or more different circRNA detection methods has been reported as highly successful in removing false positive predictions [43]. Second, since false positives in circRNA detection largely originate from repetitive sequences, we performed a BLAST search in the human and HSV-1 genomes for the 20 nt on either side of the circular junction. This confirmed that both sequences are unique. Third, we ran our alignment-based circRNA detection approach with a required overlap of at least 40 nt on either side of circular junctions on the WT 4sU-seq and total RNA-seq time-courses. This still identified up to 170 reads for the NEAT1_2 circRNA in total RNA and up to 44 in 4sU-RNA (17–36% of reads identified with a ≥10 nt overlap) and confirmed the strong induction of this circRNA (S14A and S14B Fig). Finally, we implemented a method to determine so-called “confirming read pairs”, i.e., pairs of reads where both reads could be aligned linearly to the genome within a circRNA region but only in crosswise direction that is only consistent with a circRNA but not a linear transcript (S14C Fig). We selected those confirming read pairs for which their genomic distance would imply a fragment size ≥500 nt if they originated from a linear transcript and the fragment size would be smaller if they originated from the circRNA. In this way, we identified up to 45 confirming read pairs per replicate in WT infection for hsa_circ_0003812 (S14B Fig). NEAT1_2 circular splicing was accompanied by linear splicing both within and downstream of the circRNA region but not within the NEAT1_1 region (Fig 3B and 3C, S13 and S15 Figs). None of these linear splicing events were annotated, but some were already observed at very low rates in uninfected cells. As read counts for the most frequent linear splice sites were comparable or even higher than for the induced hsa_circ_0003812 circular junction, NEAT1_2 linear splicing was also increased in absolute levels. Calculation of splicing rates (= number of junction reads / number of exon-intron reads crossing the corresponding acceptor and donor splice sites) showed that linear splicing was indeed induced and not simply observed more frequently due to the previously reported up-regulation of NEAT1 transcription in HSV-1 infection [29, 31] (Fig 3D). As linear splicing was also induced downstream of identified circRNAs and several of the confirming read pairs for the hsa_circ_0003812 circRNA included linear splice junctions, both linear and circular NEAT1_2 transcripts were further spliced in HSV-1 infection. Canonical GT-AG splice signals were used both for the circular and the most frequent linear splice junctions (see Fig 3E for a diagram) and they tended to be conserved at least among primates (S16 Fig). Since circRNA biogenesis by back-splicing requires the spliceosome [11], our data suggest that circular and linear splicing of NEAT1_2 are linked, with the circRNA originating from back-splicing during NEAT1_2 linear splicing (see also Fig 1A). Splicing rates and their increases compared to mock differed substantially between the different linear splice junctions within the circRNA region (Fig 3D), indicating presence of alternative splicing isoforms of the circRNA. Induction of linear splicing was also observed in 4sU-, chromatin-associated and nucleoplasmic RNA and in ΔICP27 and Δvhs infection (Fig 3C and 3D, S15 Fig), although it was less pronounced in ΔICP27 infection. In summary, both linear and circular splicing of NEAT1_2 is induced during HSV-1 infection while the RNA is still associated with the chromatin. Although ICP27 is not required for NEAT1_2 splicing, it may contribute. Recently, we also performed total RNA-seq of infection with mock, HSV-1 strain F (WT-F) and its ΔICP22 mutant at 8 and 12 h p.i. (n = 2) [57]. We furthermore generated telomerase-immortalized human foreskin fibroblasts (T-HFs) that express either ICP22 in isolation (T-HFs-ICP22 cells) or in combination with ICP27 (T-HFs-ICP22/ICP27 cells) upon doxycyclin (Dox) exposure. Total RNA-seq of T-HFs-ICP22 and T-HFs-ICP22/ICP27 cells was performed both with and without Dox exposure (n = 2). Ectopic co-expression of ICP22 and ICP27 lead to induction of both circular and linear splicing of NEAT1_2 (Fig 4A). However, ICP22 knockout did not abolish HSV-1-induced NEAT1_2 circular and linear splicing (Fig 4A, S17 Fig), thus ICP22 expression is not required. This also confirmed induction of linear and circular NEAT1_2 splicing in a third HSV-1 strain. Although circRNA reads were recovered after Dox-induced expression of ICP22, this was limited to 2 reads in one replicate (Fig 4B). To investigate whether ICP27 expression alone could induce NEAT1_2 splicing, we analyzed our recently published 4sU-seq data of mock and WT KOS infection and of cells transfected with an ICP27-expressing plasmid [34]. This again confirmed induction of both linear and circular NEAT1_2 splicing in WT KOS infection but recovered only one circular read upon ICP27 overexpression in one replicate (Fig 4C). Notably, NEAT1 was upregulated upon ectopic co-expression of ICP22 and ICP27 (2.2-fold, p<10−11) and in ΔICP22 (2.1-fold, p<10−4) and ΔICP27 (4.9-fold, p<10−12) infection, but not upon expression of either ICP22 or ICP27 alone. In summary, our results show that neither ICP27 nor ICP22 are required for induction of linear and circular NEAT1_2 splicing but co-expression of both proteins is sufficient. Considering the small number of circular reads recovered when expressing either one of these proteins alone, we can neither confidently confirm nor exclude that either ICP22 or ICP27 alone may be sufficient to (weakly) induce NEAT1_2 circular splicing.
Since NEAT1_2 is also up-regulated in IAV infection [29], we investigated presence of circRNAs in previously published total RNA-seq time-courses of influenza A/California/04/09 (H1N1), A/Wyoming/03/03 (H3N2), and A/Vietnam/1203/04 (H5N1) HALo virus infection (3, 6, 12 and 18 h p.i. plus time-matched controls) of human tracheobronchial epithelial cells (HTBE, multiplicity of infection (MOI) = 5) and monocyte-derived macrophages (MDM, MOI = 2) [41]. The H5N1 HALo mutant virus is an attenuated virus generated from WT influenza A/Vietnam/1203/04. H1N1 infection of HTBE cells included 24 h p.i. and two H5N1 infection time-courses of MDM cells were performed. This revealed a strong enrichment of NEAT1_2 circular splicing in H5N1 infection of both cell types and H3N2 infection of MDM cells as well as a weak enrichment in H3N2 infection of HTBE cells and H1N1 infection of MDM cells (Fig 5A, S18 Fig). The most highly expressed NEAT1_2 circRNA was again hsa_circ_0003812 circRNA as for HSV-1 infection (marked by a blue rectangle in Fig 5A) and its enrichment increased with the duration of infection. Similar trends were observed for linear splicing events, both within and downstream of the circRNA region (Fig 5B). Notably, both the hsa_circ_0003812 circRNA and linear splicing events were also detected in some of the time-matched controls, however normalized circRNA counts and linear splicing rates were significantly lower than in infected cells (with the exception of H1N1 infection, S18A Fig). IAV infection also induces the degradation of host RNAs via two mechanisms. Cap-snatching by the viral RNA-dependent RNA polymerase leads to cleavage of host RNAs downstream of the cap and is required for synthesis of viral mRNAs [58]. The PA-X protein produced by all IAV strains [59] selectively degrades RNAs transcribed by Pol II, including long non-coding RNAs, and is localized and acts predominantly in the nucleus [60]. To exclude that enrichment of the NEAT1_2 circRNA is simply due to degradation of linear transcripts without increased biogenesis of the circRNA, we also investigated overall enrichment of circRNAs in IAV infections using linear regression analysis as for HSV-1 infection. This indeed showed an enrichment of circRNAs by up to 1.4-fold in H1N1 infection, 6.8-fold in H3N2 infection and 4.6-fold for H5N1 infection (S19A–S19G Fig). It also revealed a much higher enrichment for several other circRNAs, such as a circRNA of the ZC3HAV1 (zinc finger CCCH-type containing, antiviral 1) gene. ZC3HAV1 is induced during influenza A/WSN/33 (WSN/H1N1) infection and ectopic expression of ZC3HAV1 inhibits WSN replication [61]. Thus, the ZC3HAV1 circRNA is likely increasingly produced during IAV infection due to up-regulation of ZC3HAV1 transcription and associated increased circRNA biosynthesis and is not only enriched due to linear host RNA degradation. We compared fold-changes in normalized circRNA counts between infection and time-matched controls for 7077 well-expressed circRNAs (normalized circRNA count >1 in at least one condition), which showed that the hsa_circ_0003812 NEAT1_2 circRNA was among the 1% most enriched circRNAs in H3N2 and H5N1 infection of MDM cells (S19H Fig). Occasionally, it was even more enriched than the ZC3HAV1 circRNA. Moreover, splicing rates of linear NEAT1_2 transcripts, which should not escape host mRNA degradation in IAV infection, also increased during infection (Fig 5B). Thus, our results indicate that–at least in H3N2 and H5N1 infection–enrichment of NEAT1_2 circular splicing results from increased biogenesis of the circRNA and not only from IAV-induced degradation of host mRNAs. Since both HSV-1 and IAV infection disrupt transcription termination, we also evaluated read-through for IAV infection to correlate this to the extent of induction of NEAT1_2 splicing. Paralleling NEAT1_2 induction, read-through was higher in MDM cells than in HTBE cells and higher in H5N1 infection than in H3N2 infection (S20 Fig), while H1N1 infection led to the lowest extent of read-through. Disruption of transcription termination in IAV infection is mediated by the viral NS1 protein via inhibition of the CPSF30 subunit of the cleavage and polyadenylation specificity factor (CPSF) [62]. Our results are consistent with previous reports that the NS1 protein of A/California/04/09 H1N1 virus cannot bind and inhibit CPSF30 [63], while NS1 of the H3N2 and H5N1 strains can [64, 65]. However, a recent report indicated that the NS1-CPSF30 interaction is not necessary for disruption of transcription termination in IAV infection [40]. This could explain why read-through is still observed in H1N1 infection. A similar observation can be made in HSV-1 infection. Although HSV-1-induced disruption of transcription termination is mediated by ICP27 via interaction with CPSF subunits [34], reduced but significant read-through is also observed in ΔICP27 infection likely as a stress response [22, 39]. Due to this correlation between induction of NEAT1_2 splicing and read-through and a recent report that heat stress also up-regulates NEAT1_2 expression [38], we also analyzed our previously published 4sU-seq data of 1 and 2 h salt and heat stress regarding NEAT1_2 splicing. However, neither circular nor linear NEAT1_2 splicing was induced in either of the two stress conditions.
To shed additional light on the molecular mechanism(s) governing NEAT1_2 splicing, we searched the recount3 database [66] using Snaptron [67] for presence of the novel linear NEAT1_2 splicing events. Recount3 provides read coverage data for >700,000 publicly available human and mouse RNA-seq samples from the Sequence Read Archive (SRA). Snaptron allows rapidly searching samples in recount3 for (linear) splice junctions within specific genomic regions. Our Snaptron query recovered 6,391 human samples in which at least one of the four most frequent NEAT1_2 linear splice junctions (see Fig 3E) induced in HSV-1 infection was covered by ≥10 reads. This included our HSV-1 RNA-seq experiments, several other studies on HSV-1 infection, the IAV infection time-courses, 62 studies focusing on circRNAs and 21 studies using RNase R (S21A Fig). The latter provide further support for the association of linear and circular splicing of NEAT1_2. Particularly high numbers of NEAT1_2 splice junctions were found in samples of blood cells, such as platelets, erythrocytes, peripheral blood mononuclear cells (PBMCs), erythroleukemia (K562) cells, and more (S21A Fig). Consistently, one of these studies reported that circRNAs are strongly enriched in platelets and erythrocytes [68], which both lack a cell nucleus and thus de novo transcription. In absence of de novo transcription, the high stability of circRNAs leads to their enrichment relative to linear RNAs. Application of our circRNA alignment pipeline to two RNA-seq experiments for platelets [69] and erythrocytes [70] identified multiple NEAT1_2 circRNAs including the hsa_circ_0003812 circRNA induced by HSV-1 infection (S21B Fig). Interestingly, almost no expression was detected in the ~4 kb between the 3’end of the NEAT1_1 transcript and the 5’end of hsa_circ_0003812, however, expression extended beyond hsa_circ_0003812. This either implied novel linear transcripts beginning near the 5’end of hsa_circ_0003812 or a novel circRNA not included in circBase. Indeed, our de novo circRNA detection pipeline identified a novel circular junction connecting the 5’ end of hsa_circ_0003812 with the 3’end of the downstream expressed region. This confirms that only circular, but not linear, NEAT1_2 transcripts were present in these cells and that these circRNAs are also linearly spliced. It also shows that NEAT1_2 circRNAs naturally occur in uninfected cells, however likely at such low levels that they only become detectable upon degradation of linear transcripts. It should be repeated that NEAT1_2 has been shown to be up-regulated in HSV-1 and IAV infection by RT-qPCR [29], thus enrichment of NEAT1_2 circular splicing in HSV-1 and IAV infection cannot be explained by loss of de novo transcription followed by degradation of linear transcripts. However, this partly explains the presence of NEAT1_2 circular and linear junctions in uninfected cells from the IAV infection time-courses as MDM cells were obtained from blood. Interestingly, the Snaptron search also identified NEAT1_2 linear splicing after inhibition of CDK7 by THZ1 in multiple cancer cell lines, which was not observed in untreated cells. CDK7 is an essential part of the TFIIH transcription factor complex, phosphorylates Pol II CTD at Ser5 residues and plays key roles in maintenance of promoter-proximal Poll II pausing and the regulation of transcription elongation [71]. Overexpression of CDK7 has been associated with a number of cancers and correlated to poor prognosis [71]. THZ1 is a highly selective covalent inhibitor of CDK7 with anti-tumor activity against multiple different types of cancer [72], but also inhibits CDK12 and CDK13 at higher concentrations [73]. Experiments identified with the Snaptron search cover prostate cancer (VCaP, LNCaP and DU145 cells) [42], bladder cancer (HCV-29 cells) [74], esophageal squamous cell carcinoma (TE7 and KYSE510 cells) [75], nasopharyngeal carcinoma (C666-1, HK1 and HNE1 cells) [76], pancreatic ductal adenocarcinoma (BxPC3, MiaPaCa-2 and PANC1 cells) [77] and chordoma (UM-Chor1 cells) [78]. Further literature search identified a second study of chordoma cells (UM-Chor1 and CH22) [79] with up to 24 h THZ1 treatment and a study on THZ1 treatment of B-cell acute lymphocytic leukemia cells (Nalm6) [80]. Realignment of corresponding samples confirmed linear splicing upon THZ1 treatment in all cell lines apart from DU145 and showed that splicing rates increased with the duration and dosage of THZ1 treatment (Fig 6A and 6B and S22A Fig). Surprisingly, considering the previously observed strong association between linear and circular NEAT1_2 splicing, no NEAT1_2 circRNAs were recovered for any of these samples neither with our alignment-based nor our de novo circRNA discovery pipeline. However, in all cases almost no circRNAs (maximum 105 circRNAs, mean 18) were recovered at all, not even the most ubiquitous circRNAs such as HIPK3, CORO1C or ASXL1. As read length was ≥75 nt for all samples, this cannot be explained by short read length. It rather suggests that circRNAs were generally not or only poorly recovered in all these studies. While a literature search identified no other studies for specific CDK7 inhibition without poly(A) selection, the Snaptron search also recovered linear NEAT1_2 splicing in chromatin-associated RNA of HEK293 cells upon 2 h treatment with DRB [81]. DRB inhibits CDK7 [82] but also the CDK9 kinase component of the positive P-TEFb complex [83] and other kinases [84]. The hsa_circ_0003812 NEAT1_2 circRNA induced by HSV-1 and IAV infection was indeed observed upon DRB treatment, but not any of the other NEAT1_2 circRNAs present in platelets and erythrocytes (S22B Fig). This provides some evidence that CDK7 inhibition leads to NEAT1_2 circular splicing. In any case, identification of linear splicing in absence of circRNAs shows that the linear NEAT1_2 isoform is spliced upon CDK7 inhibition. As no circRNAs were recovered, induction of linear splicing cannot be an artefact of circRNAs being enriched after THZ1-mediated transcription inhibition. Furthermore, spliced linear NEAT1_2 transcripts should not be more or less stable than unspliced NEAT1_2 transcripts as splicing occurs upstream of the last 100 nt of NEAT1_2 that form the stabilizing helical structure. Interestingly, however, CDK7 inhibition led to a relative increase of NEAT1_2 compared to NEAT1_1, suggesting that NEAT1_2 transcription was less strongly reduced by CDK7 inhibition than NEAT1_2 transcription or potentially even increased. To exclude that inhibition of transcription per se leads to NEAT1_2 splicing, we investigated data from one of the UM-Chor1 experiments that included actinomycin D (Act-D) treatment, which inhibits transcription but not via inhibition of cyclin-dependent kinases (CDKs) [85]. Act-D treatment, while leading to a similar relative increase in the abundance of NEAT1_2 compared to NEAT1_1 as THZ1 treatment, did not induce NEAT1_2 splicing (S22A Fig). In addition, siRNA-mediated knockdown of CDK7 in VCaP and LNCaP cells also led to NEAT1_2 splicing and a relative increase in NEAT1_2 compared to NEAT1_1 (Fig 6C), including many more novel splicing events. In contrast to splicing observed in HSV-1 and IAV infection and upon THZ1 treatment, some of these splicing events extended into the NEAT1_1 region. This indicates that indeed inhibition of CDK7, and not of other CDKs, leads to NEAT1_2 splicing. Interestingly, siRNA-mediated knockdown of the MED1 component of the Mediator complex in VCaP and LNCaP cells had the same effect (Fig 6C). CDK7 directly phosphorylates MED1 at Threonine 1457 [42], which together with a phosphorylation at Threonine 1032 promotes MED1 association with the Mediator complex [86]. Mediator is responsible for communicating regulatory signals from gene-specific transcription factors to Pol II and in this way impacts transcription at multiple stages [87]. Our results suggest that the effect of CDK7 inhibition on NEAT1_2 splicing may follow from the loss of MED1 phosphorylation and thus loss of MED1 association with Mediator. Interestingly, CDK7 inhibition also induced read-through transcription, albeit to a lesser degree than HSV-1 or H5N1 infection (Fig 6D). Read-through transcription after THZ1 treatment was comparable to H1N1 or H3N2 infection of HTBE cells, most pronounced in THZ1-treated Nalm6 cells and increased with the duration and dosage of THZ1 treatment in UM-Chor1 cells. In addition, both CDK7 and MED1 knockdown induced read-through at levels comparable to THZ1-treated Nalm6 cells. Moreover, THZ1 treatment also resulted in extended antisense transcription at promoters. In summary, our results show that CDK7 inhibition impacts RNA transcription and processing at multiple levels.
In this study, we report on novel circular and linear splicing of the long NEAT1_2 isoform, which is induced in both HSV-1 and IAV infection. We focused on NEAT1_2 in this study for three reasons: (i) The NEAT1_2 circRNA was the only circRNA for which levels were clearly and massively increased in HSV-1 infection and which was not simply enriched by vhs-mediated degradation of linear RNAs. (ii) It was one of the few circRNAs observed at high levels in newly transcribed 4sU-, nucleoplasmic and chromatin-associated RNA during HSV-1 infection. In contrast to these other few circRNAs, however, the NEAT1_2 circRNA is not or rarely found in uninfected cells. This indicates high levels of de novo synthesis of this circRNA during HSV-1 infection. (iii) NEAT1_2 has not previously been reported to be spliced either in a circular or linear fashion except for one linear splicing event towards the 3’end of NEAT1_2. This splicing event represented a shorter 793 nt version of the last “intron” in Fig 3E with the same 3’ but a different 5’ splice site and was recently reported to increase extractability of NEAT1_2 from the protein phase [88]. Although it was also induced in HSV-1 and IAV infection, it was not among the most frequent splicing events. Re-alignment of the RNA-seq data by Chujo et al. confirmed this splicing event as well as splicing of the full last “intron” with very few reads (≤3), but no other splicing. As NEAT1_2 participates in many RNA-protein interactions in paraspeckles, its unspliced form can only be poorly extracted from the protein phase [88]. Similar observations were made for other architectural RNAs (arcRNAs) of nuclear bodies. As genomic deletion of the 793 nt intron identified by Chujo et al. did not significantly affect extractability, splicing itself appeared to be responsible for increased extractability. Chujo et al. concluded that NEAT1_2 as well as other arcRNAs must remain unspliced to allow formation of corresponding nuclear bodies. They hypothesized that competition by protein components of paraspeckles for binding of NEAT1_2 normally prevents association of splicing factors with NEAT1_2 and thus NEAT1_2 splicing. This raises the possibility that NEAT1_2 splicing during HSV-1 and IAV infection may simply be a by-product of NEAT1_2 up-regulation and thus increased availability of NEAT1_2 and consequently reduced competition between splicing factors and paraspeckle protein components. Reduced competition would also be consistent with the observation that HSV-1 infection does not alter levels of paraspeckle proteins [29]. Furthermore, NEAT1_2 binds to serine and arginine rich splicing factor 2 (SRSF2) in HSV-1 infection and is required for association of SRSF2 with promoters of the HSV-1 genes ICP0 and thymidine kinase (TK), which are upregulated by SRSF2 [89]. However, the SRSF2 binding sites in NEAT1_2 that are enriched upon HSV-1 infection are upstream of the HSV-1-induced circRNA. Thus, increased binding between SRSF2 and NEAT1_2 is likely not responsible for NEAT1_2 splicing. Moreover, this means that the linear NEAT1_2 transcript is required for SRSF2 association. We demonstrated that the two immediate-early HSV-1 proteins ICP22 and ICP27 were sufficient to induce NEAT1_2 circular and linear splicing. As outlined in the introduction, both proteins are known to be involved in processes that impact splicing [32, 33, 36]. However, neither of these two proteins were required nor sufficient on their own for (significant) induction of NEAT1_2 splicing. A possible explanation for this observation could be that NEAT1 was up-regulated upon co-expression of ICP22 and ICP27 and in null-mutant infections of either protein, but not upon expression of either ICP22 or ICP27 alone. Thus, up-regulation of NEAT1_2 could be the trigger for NEAT1_2 splicing rather than specific functions of ICP22 and ICP27. One line of evidence arguing against NEAT1_2 splicing being only a by-product of NEAT1_2 up-regulation is the absence of NEAT1_2 circular or linear splicing in heat stress despite NEAT1_2 also being up-regulated in the heat shock response [38]. Previously, NEAT1_2 knockdown in HeLa cells was shown to reduce HSV-1 glycoprotein density and intensity and inhibit plaque formation, an indicator of mature virus production [31]. The vital role of NEAT1_2 for HSV-1 infection was also confirmed in vivo as a thermosensitive gel containing NEAT1_2 siRNA could heal HSV-1-induced skin lesions in mice [31]. This siRNA targeted the first “exon” within the circRNA (see Fig 3E), which is contained in both linear and circular spliced and unspliced NEAT1_2 isoforms, thus it should deplete all of them. Interestingly, eCLIP data from ENCODE for the two core paraspeckle proteins SFPQ and NONO shows enriched binding of these proteins in NEAT1_2 within the two most frequently spliced “introns” (1 and 2 in Fig 3E), the “exon” between them, upstream of the circRNA 5’ end and towards the 3’end of NEAT1_2, but not in the first two “exons” and “intron” 3. Thus, splicing removes parts of the NEAT1_2 RNA that are important for binding by SFPQ and NONO, which are required for paraspeckle integrity [23]. Moreover, spatial organization of NEAT1_2 in paraspeckles is highly ordered, with 5’ and 3’ ends of NEAT1_2 confined to the periphery and its central sequences localized in the core of paraspeckles [90]. This makes it unlikely that either spliced linear or circular NEAT1_2 isoforms can provide scaffolds for paraspeckle assembly. Accordingly, up-regulation of NEAT1_2 splicing by unknown mechanisms could also be part of the immune response against HSV-1 to at least dampen paraspeckle up-regulation. On the other hand, SFPQ and NONO have recently been reported to bind around circRNA loci, and SFPQ depletion led to reduced expression of a subset of circRNAs [91]. Thus, these paraspeckle proteins itself could be involved in NEAT1_2 splicing. Since NEAT1 is up-regulated in both HSV-1 and IAV infection and NEAT1_2 plays a general antiviral role by mediating IL-8 up-regulation, NEAT1_2 splicing observed in both HSV-1 and IAV infection could be mediated by similar mechanisms or play similar roles. However, while NEAT1 is known to have a proviral function in HSV-1 as outlined above, no proviral function of NEAT1 has been reported for IAV infection. Thus, the impact of NEAT1_2 splicing in HSV-1 and IAV infection likely differs at least partly. Linear NEAT1_2 splicing, but likely also circular splicing, is also observed upon CDK7 inhibition in cancer cell lines and NEAT1_2 circRNAs are abundantly found in various blood cells in absence of de novo transcription. However, transcription inhibition alone cannot explain NEAT1_2 splicing in HSV-1 and IAV infection and upon CDK7 inhibition as NEAT1 is up-regulated in HSV-1 and IAV infection and transcription inhibition by Act-D does not induce NEAT1_2 splicing. Since CDK7 inhibition leads to a relative increase of NEAT1_2 compared to NEAT1_1, NEAT1_2 expression may not actually be substantially reduced or even increased upon CDK7 inhibition. This again raises the possibility that an increase in NEAT1_2 upon CDK7 inhibition reduces the competition between paraspeckle proteins and splicing factors, resulting in NEAT1_2 splicing. On the other hand, DRB treatment, which also inhibits CDK7 among other CDKs and leads to NEAT1_2 circular and linear splicing, induces dissociation of paraspeckle proteins from NEAT1_2 and disappearance of paraspeckles [23, 92]. Considering the likely negative impact of NEAT1_2 splicing on paraspeckle formation discussed above, NEAT1_2 splicing could thus contribute to loss of paraspeckles upon DRB treatment. However, paraspeckle disassembly is also observed upon Act-D treatment, which does not induce NEAT1_2 splicing [23]. In contrast, both HSV-1 and IAV infection were found to induce excess formation of paraspeckles, although in IAV infection they were slightly diffuse [29]. Thus, paraspeckles can be increased even in presence of NEAT1_2 splicing. Considering the substantial induction of NEAT1_2 levels in HSV-1 and IAV infection [29], levels of unspliced NEAT1_2 are likely still sufficiently increased–despite a loss to splicing–to induce excess paraspeckle formation. Alternatively, however, NEAT1_2 splicing may only occur in HSV-1 and IAV infection because NEAT1_2 is more up-regulated than paraspeckle proteins, making NEAT1_2 accessible to splicing factors. Nevertheless, it is tempting to speculate that diffuseness of paraspeckles in IAV infection could at least be partly due to altered paraspeckle formation on (partly) spliced NEAT1_2 isoforms, as NEAT1_2 was not as strongly up-regulated upon IAV infection as upon HSV-1 infection. Interestingly, we found that selective CDK7 inhibition by THZ1 also disrupts transcription termination in cancer cell lines, albeit less strongly than HSV-1 and H5N1 infection. This was surprising considering a recent report that 1 h THZ1 treatment suppresses Pol II read-through at gene 3’ends in acute myeloid leukemia [93]. It is, however, consistent with earlier reports of impaired transcription termination and 3′-end processing of an snRNA, a polyadenylated mRNA and a histone RNA upon CDK7 inhibition [94, 95]. Our results thus demonstrate for the first time widespread disruption of transcription termination upon CDK7 inhibition. Since read-through transcription was only observed upon long-term or high-dose THZ1 treatment or CDK7 knockdown, a possible explanation for the discrepancy to the study by Sampathi et al. is that disruption of transcription termination upon CDK7 inhibition is a downstream response. Interestingly, the relative increase in expression of the non-polyadenylated NEAT1_2 isoform compared to the polyadenylated NEAT1_1 isoform observed upon CDK7 inhibition was also observed in HSV-1 and IAV infection, although not consistently. This raises the possibility that read-through transcription of the NEAT1_1 poly(A) site may be involved in induction of NEAT1_2 splicing by increasing abundance of NEAT1_2. Although salt and heat stress did not induce NEAT1_2 splicing, this does not exclude a role of read-through transcription. While there are strong overlaps between the genes affected by read-through in HSV-1 infection and stress conditions, there are also clear context- and condition-specific differences [22]. Furthermore, the mechanisms underlying disruption of transcription termination in stress conditions remain elusive. This contrasts with HSV-1 and IAV infection, where the HSV-1 ICP27 protein and the IAV NS1 protein have been shown to disrupt transcription termination via interaction with CPSF subunits [34, 41, 62]. It is important to note that our analyses of NEAT1_2 splicing and read-through upon CDK7 inhibition were performed in cancer cell lines, which are particularly susceptible to CDK7 inhibition [72]. Dysregulation of alternative polyadenylation, often associated with wide-spread shortening of 3’UTRs, is common in cancers [96]. NEAT1 also plays different roles in cancer development and both NEAT1 isoforms have been proposed as cancer biomarkers [97], with NEAT1_2 considered a tumor suppressor and NEAT1_1 considered to be oncogenic [98]. Thus, read-through, relative increase of NEAT1_2 expression, or NEAT1_2 splicing upon CDK7 inhibition could be linked to susceptibility of cancer cell lines to CDK7 inhibition. Our study identified several parallels between HSV-1 and IAV infection, on the one hand, and CDK7 inhibition, on the other hand. This not only includes read-through transcription and NEAT1_2 splicing, but also induction of antisense transcription upon CDK7 inhibition. We previously reported on widespread activation of antisense transcription in HSV-1 infection [99]. Moreover, widespread loss of promoter-proximal pausing was recently reported both upon CDK7 inhibition [93, 100] and HSV-1 infection [101]. Previously, LDC4297A, a different selective inhibitor of CDK7, has been shown to have strong antiviral activity against HSV-1 as well as other herpesviruses (but only low efficacy against H1N1) [102] and CDK7 has been found to be a target of CDK inhibitors inhibiting HSV replication [103]. Both studies confirm a role of CDK7 in HSV-1 infection. It is tempting to speculate that recruitment of CDK7-containing complexes, similar to recruitment of elongation factors like FACT by ICP22 [104], to the HSV-1 genome may lead to a depletion of CDK7 on host genes, thus mimicking the effects of CDK7 inhibition. In summary, our findings highlight potential important roles of NEAT1_2 splicing and CDK7 in HSV-1 and/or IAV infection.
All RNA-seq data analyzed in this study were downloaded from the SRA. SRA project IDs: HSV-1 WT infection time-course: SRP044766; HSV-1 Δvhs infection time-course: SRP192356; RNA-seq of subcellular fractions: SRP110623, SRP189489, SRP191795 (same experiment but data for mutant viruses submitted separately); total RNA-seq for mock and WT KOS infection with ERCC spike-ins: SRP321121 (samples: SRR14632002, SRR14631995); T-HFs-ICP22 and T-HFs-ICP22/ICP27 cells, WT strain F and ΔICP22 infection: SRP340110; WT KOS, ΔICP27 and ICP27 overexpression: SRP189262; IAV infection time-courses: SRP091886, SRP103821; platelets: ERP003815; erythrocytes: SRP050333; CDK7 inhibition/knockdown: VCaP, LNCaP and DU145 cells: SRP179971; HCV-29 cells: SRP217721; TE7 and KYSE510 cells: SRP068450; C666-1, HK1 and HNE1 cells: SRP101458; BxPC3, MiaPaCa-2 and PANC1 cells: SRP165924; UM-Chor1 and CH22 cells: SRP166943, SRP270819; Nalm6 cells: SRP307127; DRB treatment of HEK293 cells: SRP055770. RNase R treatment: SRP197110, SRP152310.
Sequencing reads were downloaded from SRA using the sratoolkit version 2.10.8 and aligned against the human genome (GRCh37/hg19) and human rRNA sequences using ContextMap2 version 2.7.9 [105] using BWA as short read aligner [50] and allowing a maximum indel size of 3 and at most 5 mismatches. For HSV-1 infection RNA-seq data, alignment also included the HSV-1 genome (Human herpesvirus 1 strain 17, GenBank accession code: JN555585). For the two repeat regions in the HSV-1 genome, only one copy was retained each, excluding nucleotides 1–9,213 and 145,590–152,222 from the alignment. SAM output files were converted to BAM files using samtools [106]. Following read mapping, fragment size distribution was determined from BAM files using the Picard CollectInsertSizeMetrics tool [107]. The linear RNA-seq mapping was used for identification of linear splice sites, filtering of identified circRNA reads (see below) and visualization of read coverage on genes.
For de novo detection of circRNAs, we applied CIRI2 [45] and circRNA_finder [44] in parallel to all reads for each sample (including both reads aligned and unaligned in the linear RNA-seq mapping) (outline in Fig 2A). Combination of two complementary circRNA detection algorithms, such as CIRI2 and circRNA_finder, has been recommended to remove algorithm-specific false positives commonly observed in circRNA detection [43]. Both CIRI2 and circRNA_finder allow de novo detection of circRNAs using splice sites not annotated in the human genome. CIRI2 is based on identifying so-called paired chiastic clipping (PCC) signals in BWA alignments, which are pairs of clipped read alignments (i.e., local alignments for different substrings of the read) where a downstream part of the read aligns upstream of an upstream part of the read. Such reads are denoted as back-spliced junction (BSJ) reads and are then further filtered by CIRI2 to remove false positives (see Fig 2A and the original publication for details). For this purpose, we first aligned all reads using BWA with parameters recommended by the developers of CIRI2 (= default BWA parameters, except for the -T option (= minimum score to output an alignment), which was set to 19). With default BWA options, matches in alignments are given a score of 1, mismatches a penalty of 4, the gap open penalty was 6 and the gap extend penalty was 1. As a consequence, mismatches on read segments flanking each side of the circular junction are only allowed if corresponding read segments are ≥ 24 nt. circRNA_finder is based on filtering chimeric alignments obtained with STAR [49], in which different parts of a read align to two distinct regions of the genome in a manner not consistent with “normal” linear transcripts. We thus first aligned all reads with STAR using the default parameters used by the developers of circRNA_finder, i.e., chimSegmentMin (minimum total length of the chimeric segment) = 20, chimScoreMin (minimum total score of the chimeric segments) = 1, alignIntronMax (maximum intron length) = 100000, outFilterMismatchNmax (maximum number of mismatches per pair) = 4, alignTranscriptsPerReadNmax (maximum number of different alignments per read to consider) = 100000, outFilterMultimapNmax (maximum number of loci the read is allowed to map to) = 2. circRNA_finder then further filters chimeric alignments to identify circRNAs and remove false positives (see Fig 2A and the original publication for details). CircRNAs were only further analyzed if they were detected independently by both algorithms. Circular reads identified by either algorithm were then pooled and subsequently all reads were removed that could be mapped in a linear fashion to the genome using ContextMap2. Only circRNAs with at least two supporting reads remaining in the same sample were further analyzed. Fragment lengths were not explicitly considered for circRNA detection, but all used algorithms have different criteria regarding the allowed distances and positions of the two reads in a read pair relative to each other (see original publications for details). The complete pipeline is available as a workflow for the workflow management system Watchdog [108] in the Watchdog workflow repository (https://github.com/watchdog-wms/watchdog-wms-workflows/tree/master/circRNA_Detection).
Candidate circular junction sequences were generated from putative spliced circRNAs sequences downloaded from circBase [48]. RNA-seq reads were aligned against circular junction sequences using BWA [50] with default parameters (i.e., -T is set to 30). BWA clips a read, i.e., outputs only a local alignment of the read, if the best local alignment score of the read minus a clipping penalty (default = 5) is at least as good as the best global alignment score. Thus, if a read only aligns on one side of the circular junction, only a local alignment to that side of the junction will be generated by BWA with at most a very short overlap to the other side of the junction with few mismatches. Reads were retained if they aligned to at least X nt (default X = 10) on either side of the junction and could not be mapped in a linear fashion anywhere to the genome using ContextMap2. In case of paired-end sequencing, the second read in the pair also had to align consistently within the circRNA region. circRNAs resulting from repetitive regions were discarded. The complete pipeline is outlined in S4 Fig.
ContextMap2 is also capable of aligning read pairs, where each individual read can be aligned linearly to the genome, but the two reads are aligned in the wrong orientation relative to each other (= crosswise, see S14C Fig). Such crosswise alignments of read pairs are inconsistent with linear transcripts and were considered a confirming read pair for a circRNA if (i) both reads were aligned within the circRNA region with a genomic distance implying a fragment size ≥ 500 nt if they originated from a linear transcript and (ii) the fragment size would be smaller if they originated from the circRNA. This fragment size cutoff was based on the observed distribution of fragment sizes, which showed that <3.7% of fragments exhibited a fragment size ≥ 500 nt.
Number of read fragments per gene and in 5kb windows downstream of genes were determined from read alignments using featureCounts [109] and gene annotations from Ensembl (version 87 for GRCh37/hg19) [110]. For strand-specific RNA-seq, the stranded mode of featureCounts was used. All fragments (read pairs for paired-end sequencing or reads for single-end sequencing) overlapping exonic regions on the corresponding strand by ≥25 nt were counted for the corresponding gene. Downstream transcription for a gene was calculated as previously described [22] as the FPKM (Fragments Per Kilobase Million) in the 5kb windows downstream of genes divided by the gene FPKM multiplied by 100. %Read-through transcription was quantified as the difference in downstream transcription between infected/treated and uninfected/control cells, with negative values set to zero.
Sashimi plots were generated with ggsashimi [111]. Heatmaps were generated with the heatmap.2 and pheatmap functions in R. Differential exon usage analysis was performed with the R Bioconductor package DEXSeq [112]. Log2 fold-changes in gene expression for WT and Δvhs infection time-courses were obtained from our recent study [18]. In this study, we used DESeq2 [113] to compare gene expression between mock and each time-point of infection for 4,162 genes without read-in transcription originating from disrupted transcription termination for an upstream gene. Genes with read-in transcription were excluded from the analysis as read-in transcription can be mistaken for induction of gene expression.
Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. | true | true | true |
||
PMC9592229 | Peng Zhang,Luhao Liu,Weiting Zhang,Jiali Fang,Guanghui Li,Lei Zhang,Jiali Li,Xuanying Deng,Junjie Ma,Kun Li,Zheng Chen | Effects of Long Noncoding RNA HOXA-AS2 on the Proliferation and Migration of Gallbladder Cancer Cells | 17-10-2022 | To explore the function and mechanism of lncRNA HOXA-AS2 in cancer-associated fibroblasts (CAFs)-derived exosomes in gallbladder cancer metastasis, and provide new research targets for the treatment of gallbladder cancer. At the same time, in order to clarify the early predictive value of lncRNA HOXA-AS2 for gallbladder cancer metastasis, and to provide a theoretical basis for clinical individualized treatment of gallbladder cancer. Methods. In our previous work, we used TCGA database analysis to find that lncRNA HOXA-AS2 was highly expressed in gallbladder cancer tissues compared with normal tissues. In this study, the expression levels of HOXA-AS2 in gallbladder cancer cell lines and control cells were first verified by QPCR and Western blot methods. Then, lentiviral tools were used to construct knockdown vectors (RNAi#1, RNAi#2) and negative control vectors targeting two different sites of HOXA-AS2, and the vectors were transfected into NOZ and OCUG-1 cells, respectively. Real-time PCR was used to detect knockdown efficiency. Then, the effects of silencing HOXA-AS2 on the proliferation, cell viability, cell migration, and invasion ability of gallbladder cancer cells were detected by MTT, plate cloning assay, Transwell migration chamber assay, and Transwell invasion chamber assay. Finally, the interaction between HOXA-AS2 and miR-6867 and the 3′UTR of YAP1 protein was detected by luciferase reporter gene. The results showed that the expression level of HOXA-AS2 in gallbladder cancer cell lines was higher than that in control cells. The expression of HOXA-AS2 in gallbladder carcinoma tissues was significantly higher than that in adjacent tissues (p < 0.05). After successful knockout of HOXA-AS2 by lentiviral transfection, the expression of HOXA-AS2 in gallbladder cancer cell lines was significantly decreased. Through cell proliferation and plate clone detection, it was found that silencing HOXA-AS2 inhibited cell proliferation and invasion. Through software prediction and fluorescein reporter gene detection, it was found that HOXA-AS2 has a binding site with miR-6867, and the two are negatively correlated, that is, the expression of miR-6867 is enhanced after the expression of HOXA-AS2 is downregulated. And the 3′UTR of YAP1 protein in the Hippo signaling pathway binds to miR-6867. Therefore, HOXA-AS2 may affect the expression of YAP1 protein by regulating miR-6867, thereby inhibiting the Hippo signaling pathway and promoting the proliferation and metastasis of gallbladder cancer cells. HOXA-AS2 is abnormally expressed in gallbladder cancer cells. HOXA-AS2 may promote the migration and invasion of gallbladder cancer cells by regulating the Hippo signaling pathway through miR-6867. HOXA-AS2 may serve as a potential diagnostic and therapeutic target for gallbladder cancer in clinic. | Effects of Long Noncoding RNA HOXA-AS2 on the Proliferation and Migration of Gallbladder Cancer Cells
To explore the function and mechanism of lncRNA HOXA-AS2 in cancer-associated fibroblasts (CAFs)-derived exosomes in gallbladder cancer metastasis, and provide new research targets for the treatment of gallbladder cancer. At the same time, in order to clarify the early predictive value of lncRNA HOXA-AS2 for gallbladder cancer metastasis, and to provide a theoretical basis for clinical individualized treatment of gallbladder cancer. Methods. In our previous work, we used TCGA database analysis to find that lncRNA HOXA-AS2 was highly expressed in gallbladder cancer tissues compared with normal tissues. In this study, the expression levels of HOXA-AS2 in gallbladder cancer cell lines and control cells were first verified by QPCR and Western blot methods. Then, lentiviral tools were used to construct knockdown vectors (RNAi#1, RNAi#2) and negative control vectors targeting two different sites of HOXA-AS2, and the vectors were transfected into NOZ and OCUG-1 cells, respectively. Real-time PCR was used to detect knockdown efficiency. Then, the effects of silencing HOXA-AS2 on the proliferation, cell viability, cell migration, and invasion ability of gallbladder cancer cells were detected by MTT, plate cloning assay, Transwell migration chamber assay, and Transwell invasion chamber assay. Finally, the interaction between HOXA-AS2 and miR-6867 and the 3′UTR of YAP1 protein was detected by luciferase reporter gene. The results showed that the expression level of HOXA-AS2 in gallbladder cancer cell lines was higher than that in control cells. The expression of HOXA-AS2 in gallbladder carcinoma tissues was significantly higher than that in adjacent tissues (p < 0.05). After successful knockout of HOXA-AS2 by lentiviral transfection, the expression of HOXA-AS2 in gallbladder cancer cell lines was significantly decreased. Through cell proliferation and plate clone detection, it was found that silencing HOXA-AS2 inhibited cell proliferation and invasion. Through software prediction and fluorescein reporter gene detection, it was found that HOXA-AS2 has a binding site with miR-6867, and the two are negatively correlated, that is, the expression of miR-6867 is enhanced after the expression of HOXA-AS2 is downregulated. And the 3′UTR of YAP1 protein in the Hippo signaling pathway binds to miR-6867. Therefore, HOXA-AS2 may affect the expression of YAP1 protein by regulating miR-6867, thereby inhibiting the Hippo signaling pathway and promoting the proliferation and metastasis of gallbladder cancer cells. HOXA-AS2 is abnormally expressed in gallbladder cancer cells. HOXA-AS2 may promote the migration and invasion of gallbladder cancer cells by regulating the Hippo signaling pathway through miR-6867. HOXA-AS2 may serve as a potential diagnostic and therapeutic target for gallbladder cancer in clinic.
Clinically, gallbladder cancer (GBC), as one of the nauseating tumors, has a very high mortality rate despite its low incidence [1], usually found in the biliary system [2, 3]. However, due to the limited potential for curative resection and its resistance to chemotherapeutic agents, gallbladder carcinoma is an aggressive malignancy with high mortality [4]. Finding therapeutic targets for gallbladder cancer is an important process to prolong the survival of patients with gallbladder cancer. Local invasion and distant metastasis are important biological features of gallbladder cancer. The development of effective molecular targets plays an important role in inhibiting the metastasis of gallbladder cancer. Many previous studies have found that miRNAs are involved in the metastatic process of gallbladder cancer. Bao et al. found that miR-101 inhibits the metastasis of gallbladder cancer [5]. MicroRNA-135a [6] and miR-20 [7] have been shown to be closely related to gallbladder cancer metastasis. In the field of oncology research, a number of studies in recent years have demonstrated that lncRNAs are involved in the formation and development of tumors [8–10]. lncRNAs in the invasion of gallbladder cancer has so far been unclear. On the basis of previous studies, screening gallbladder cancer cell lines with different metastatic characteristics and searching for differentially expressed genes/lncRNAs by sequencing is a more effective research method. Through this experimental method, Wang et al. successfully demonstrated that CLIC1 promotes the migration and invasion of gallbladder cancer cells [11]. Related studies have shown that lnc-H19 promotes gallbladder cancer metastasis by regulating EMT [12]. lnc-CCAT1 promotes gallbladder cancer metastasis by negatively regulating miR-218-5p [13]. Therefore, it is necessary to screen lncRNAs related to gallbladder cancer metastasis by lncRNA chip. In our previous work, we collected tumor tissues from patients with gallbladder cancer, isolated CAFs-derived exosomes, and used lncRNA microarray chip and TargetScan software to analyze the differential lncRNAs related to metastasis. Finally, an lncRNA with significant differential expression was screened out, namely, lncRNA HOXA-AS2. The lncRNA HOXA-AS2 is an unknown lncRNA, and its biological function and mechanism of action are unclear. Therefore, lncRNA HOXA-AS2-specific siRNAs targeting different targets which were selected for a series of experimental studies to study their effects invasion of gallbladder cancer cells at the cellular level and to explore new targets for molecular targeted therapy of gallbladder cancer.
Human gallbladder cancer cell lines: GEC, SGC-996, EH-GB, NOZ, GBC-SD, and OCUG-1. Clinical tissue samples from 15 patients with gallbladder cancer who underwent cholecystectomy for gallbladder cancer in the Second Affliated Hospital of Guangzhou Medical University and Zhujiang Hospital of Southern Medical University from May 2019 to March 2021 were collected, including cancer tissues and normal paracancerous tissues of different TNM stages. Collected clinical tissue samples were kept in liquid nitrogen until total RNA extraction.
For the extraction of total RNA from gallbladder cancer cell lines and clinical gallbladder cancer tissue samples, the specific steps refer to previous studies [14]. The extracted total RNA was synthesized into cDNA using a reverse transcription system kit (Thermo Fisher Scientific) as a reaction template for real-time fluorescence quantitative PCR. The content of lncRNA HOXA-AS2 was detected using a fluorescence quantitative PCR kit (Applied Biosystems, USA). Refer to previous studies for specific steps [15].
The NOZ cells and OCUG-1 cells in logarithmic growth phase in which the HOXA-AS2 gene was knocked out were collected and counted. Select an appropriate cell density for passage in a 96-well plate (about 5,000 cells per well), set 3 parallel wells, and take out a well every 48 hours. Add preprepared MTT solution (Aladdin, Shanghai) and DESO solution.
The gallbladder cancer cell lines before and after the knockout of the HOXA-AS2 gene were selected for cell clone formation experiments, including, trypsinizing and counting cells in logarithmic growth phase, and inoculating cells with appropriate density in 6-well plates (each well). Seed about 500-2000 cells, mix well, and set 3 parallel wells. Finally, observe under an inverted microscope, count and take pictures, and make statistics of the results.
Select NOZ and OCUG-1 cells successfully transfected with RNAi#1, RNAi#2, and NC (and the cells are in logarithmic growth phase), and use serum-free cell culture medium to culture the cells overnight before the experiment to reduce the effect of serum on the experiment. Cells were then trypsinized, washed 3 times with serum-free medium, counted, and made into suspension. Add the cell suspension to the Transwell chamber and incubate the cells with serum-free medium. PBS solution was used to wash the cells that did not invade the upper layer and were observed, photographed, and counted under a microscope (Zeiss, Germany). Data processing and result analysis are then carried out. The specific steps of related experiments refer to previous studies [16].
Select freshly grown 30% monolayer cells as transfected cells for future use. The transfection groups are as follows: blank group NOZ cells, OCUG-1 cells (without any treatment); negative control NC (transfected with Scramble siRNA); and experimental group: against specific valid sequences of lncRNA HOXA-AS2 targeting two different target sites and GV112 lentiviral integration plasmids (RNAi#1, RNAi#2). Change the cell culture medium to serum-free medium before infection, and add HOXA-AS2 to interfere with lentivirus sh-RNAi#1, sh-RNAi#2, and negative control lentivirus sh-Ctrl according to 10MOI (multiplicity of infection). To infect cells, add a certain amount of polybrene solution to the cell culture medium to improve the efficiency of virus infection. Normal cell passaging was performed after cells were confluent. On the 4th day after infection, the virus infection of cells was checked with an inverted fluorescence microscope, and finally NOZ and OCUG-1 cells with knockdown of HOXA-AS2 were obtained and analyzed by qRT-PCR. Interference efficiency at different sites in HOXA-AS2 was obtained.
The total protein extraction kit (Teyebio, Shanghai, China) was used to lyse and extract tissue proteins and cellular proteins. The specific experimental operation can refer to the previous research [17]. The protein concentration was subsequently determined using the BCA method (Teyebio, Shanghai, China). After the denatured protein was separated by gel running, the resultant was blotted onto a polyvinylidene fluoride membrane. The entire transfer system was placed in an ice-water mixture, and the membrane was transferred for about an hour under the conditions of 100 V, 400 mA. This was followed by overnight incubation with 5% nonfat dry milk in blocking solution. Use the desired antibody as the primary antibody to incubate the blocked PVDF membrane according to the instructions, and add an appropriate amount of secondary antibody and incubate with shaking at room temperature. A development kit (Teyebio, Shanghai, China) visualized the bands. The antibodies used are as follows: CyclinD1, p21, MMP9, snail, YAP1, p-YAP, TAZ, p-TAZ, GAPDH antibodies (1 : 2000, Abcam), and HRP labeled IgG antibody (1 : 10000, Cell Signaling Technology).
Wild and mutant HOXA-AS2/YAP1 was cotransfected with miR-6867-5p mimic/NC into HEK-293 T cells. Then luciferase activity was measured on a luciferase reporter system (Promega) using a dual-luciferase reporter gene detection kit (Beyotime, Shanghai, China).
Statistical software SPSS22.0 was used for data analysis. All data were repeated at least 3 times. The two-tailed student's t-test was used to assess the difference between the two groups, and the level of statistical difference was expressed as p value: ∗, p value <0.05.
In the TCGA database, the expression of HOXA-AS2 was analyzed in normal tissues and gallbladder cancer tissues, and it was found that HOXA-AS2 was abnormally expressed in gallbladder cancer tissues (Figure 1(a)). Analyze the level of lncRNA HOXA-AS2 in clinical tissue samples of different stages of gallbladder cancer. The expression of HOXA-AS2 in the clinical tissues of different stages of gallbladder cancer was higher than that in the corresponding normal tissues (Figure 1(b)). In addition, we further detected the expression of lncRNA HOXA-AS2 in hepatoma cells. As shown in Figure 1(c), the level of HOXA-AS2 in normal gallbladder cell GECs was used as a reference. HOXA-AS2 was highly expressed in multiple gallbladder cancer cell lines.
NOZ and OCUG-1 cells were transfected with specific shRNAs (RNAi#1, RNAi#2) targeting two different sites of lncRNA HOXA-AS2 carried by lentiviral tools, 96 hours after transfection, the expression level of RNAi#2 was only 2.6% of the Lv-shCon (NC) group (Figure 2(a)). The above results suggested that Lv-shHOXA-AS2 specifically knocked down HOXA-AS2 in NOZ and OCUG-1 cells. MTT cell viability assays and cell colony formation experiments were performed. As shown in Figure 2(b), we found that the absorbance at 490 nm of NOZ and OCUG-1 cells infected with sh-HOXA-AS2 lentivirus was significantly lower than that of cells infected with sh-Ctrl lentivirus, indicating that knockdown of SNHG16 inhibited viability of NOZ and OCUG-1 cells. Clonogenic assays showed decreased cell growth in NOZ and OCUG-1 cells knocked down HOXA-AS2 (Figure 2(c)). The results of in vitro proliferation experiments showed that the level of HOXA-AS2 was downregulated in NOZ and OCUG-1 cells and inhibited cell proliferation. In addition, in the Transwell migration and invasion experiments, we found that under the premise of maintaining the same initial cell number, after 48 hours of cell culture, the NOZ and OCUG-1 cells in the HOXA-AS2 expression-decreased group showed reduced migration and invasion cells (Figure 2(d)), suggesting that the mutation of HOXA-AS2 inhibits the migration and invasion of cancer cells in vitro.
The expression of key cell cycle-related regulators CyclinD1 and P21 proteins was detected by real-time PCR and Western blot (Figures 3(a) and 3(b)). The expression of CyclinD1 protein was found to be decreased in the RNAi#1 and RNAi#2 groups. The levels of MMP9 and Snail, which affect cell migration and invasion were subsequently detected, and the levels of MMP9 and Snail were significantly reduced after silencing lnc-HOXA-AS2 (Figures 3(a) and 3(b)).
RNA from clinical tissue samples of gallbladder carcinoma was detected using RT-PCR, and the expression of HOXA-AS2 was found to be inversely correlated with MiR-6867-5p (Figure 4(a)). Subsequently, the results were analyzed by TargetScan software, and it was found that HOXA-AS2 and miR-6867 have the same binding site (Figure 4(b)). MiR-6867-5p was significantly increased in NOZ and OCUG-1 cells in RNAi#1 and RNAi#2 group which HOXA-AS2 was knocked down (Figure 4(c)). HOXA-AS2 expression was decreased after overexpression of miR-6867 in NOZ and OCUG-1 cells and increased upon addition of miR-6867 inhibitor (Figure 4(d)). These experimental results indicated that HOXA-AS2 was inversely correlated with the expression of miR-6867-5p in related cell lines. These further support the idea that miR-6867-5p is the target of lncRNA HOXA-AS2.
We performed analysis using TargetScan prediction software and found that YAP1 was a target of miR-6867-5p (Figure 5(a)). Subsequently, by RT-qPCR and Western blot detection, the level of YAP1 was increased in the RNAi#1/RNAi#2 group and miR-6867-5p overexpression group. RNAi#1/RNAi#2 group and miR-6867-5p significantly decreased in the inhibitor group (Figures 5(b) and 5(c)). As lncRNA HOXA-AS2 was not silenced resulting in upregulation of miR-6867-5p, this may further upregulate YAP1 levels. Furthermore, we found that the luciferase activity was significantly reduced in the HOXA-AS2-miR-6867-5p group at two specific sites (Figure 5(d)). All the results demonstrate a consistent axis of regulatory relationship between lncRNA HOXA-AS2-miR-6867-5p–YAP1. The transcriptional coactivator YAP/TAZ in the Hippo signaling pathway loses its transcriptional activity when phosphorylated, and YAP/TAZ itself is a transcriptional coactivator that cannot bind DNA, so it needs to be combined with other transcription factors such as TEAD1-4, coinitiated the transcription of downstream genes. Therefore, we detected the phosphorylation and activation of YAP and TAZ after HOXA-AS2 knockdown by Western blot. The phosphorylation of YAP was significantly reduced in RNAi#1\RNAi#2 gallbladder cancer cells (Figure 6(a)). At the same time, the results of luciferase activity detection showed that the activities of TEAD1-4 transcription factors that interacted with YAP also decreased correspondingly (Figure 6(b)), indicate that silencing of HOXA-AS2 affects the expression of Hippo signaling pathway-related regulators. In addition, adding an agonist of the Hippo signaling pathway to HOXA-AS2 knockdown gallbladder cancer cells, and through cell cloning and invasion experiments found that the increased expression of Hippo signaling pathway-related regulators inhibited the proliferation and invasion of gallbladder cancer cells (Figures 6(c) and 6(d)).
Gallbladder cancer is a pathogenic malignancy, affecting 2.5 per 100,000 people [18, 19]. lncRNA HOXA-AS2 has been found to be aberrantly expressed in a variety of human tumor tissues and cells (Supplementary Figure 1). HOXA-AS2 was highly expressed in gallbladder tumor tissues and cells. The molecular occurrence and progression of tumors is an extremely complex problem, in which cell cycle disturbance and epithelial-mesenchymal transition are common features of many types of human malignant tumor cells. Cell cycle disorders lead to uncontrolled cell proliferation, which greatly enhances the ability of cells to proliferate, while epithelial-mesenchymal transition makes cells lose contact inhibition and can move around [20, 21]. Epithelial-mesenchymal transition (EMT) also plays a significant role in the migration of gallbladder cancer cells. On the basis of this study, whether EMT is involved in the regulation of AS2 on the migration of gallbladder cancer cells can be further explored in the future. Our study confirmed that reducing HOXA-AS2 expression attenuated tumor cell proliferation in gallbladder cancer cell lines by the results of MTT assay and clone formation assay. The results of in vitro cell function experiments showed that interfering with the expression of HOXA-AS2 could inhibit the speed of in vitro migration of gallbladder cancer cells and the ability to invade other cells. At the same time, Western blot analysis of the expression of regulatory factors related to cell cycle and metastasis further confirmed that downregulation of HOXA-AS2 inhibited the migration of gallbladder cancer cells, but P21 protein and CyclinD1 protein were related to cell cycle. This may be related to the antiapoptotic effect of P21 protein [22–24]. Therefore, in this study, we explored the effect of HOXA-AS2 on cell cycle by detecting the expression of P21 protein and CyclinD1 protein. Usually, lncRNAs and microRNAs (miRNAs) work together to regulate each other. We found that miR-6867-5p may be targeted by lncRNA HOXA-AS2 through bioinformatics tools. miR-6867-5p is a newly discovered miRNA with less research in the field of cancer. miR-6867-5p was found to be important in promoting angiogenesis under hypoxic conditions [25]. In this study, the expression of lncRNA HOXA-AS2 and miR-6867-5p strongly proves that lncRNA HOXA-AS2 and miR-6867-5p interact in the development of gallbladder cancer. According to the prediction of the related software, YAP1 was predicted to be the target of miR-6867-5p, which was also confirmed by the detection of the luciferase reporter gene. The core components of signal transduction, Lats1/2, and Mst1/2, are often downregulated in some tumors due to the hypermethylation of their promoters, thereby promoting the malignant transformation of tumors [26]. Yes-associated proteins (YAPs) typically regulate signal transduction and gene transcription in cells [27, 28]. YAP is generally highly expressed in many tumors, and its nuclear activity is significantly enhanced, thereby inducing tumor progression [29–31]. lncRNA GAS5 interacts with YAP1 phosphorylation and degradation to inhibit rectal cancer progression [32]. However, how the Hippo-Yap signaling pathway plays its regulatory role in gallbladder cancer remains to be further clarified. The effect of lncRNA HOXA-AS2 on gallbladder cancer cells may affect the downregulation of miR-6867-5p and further affect the role of YAP1. In our study, YAP1 was first investigated in HOXA-AS2-knockdown gallbladder cancer cells. We found that YAP1 was downregulated in miR-6867-5p inhibitor and lncRNA HOXA-AS2 unsilenced groups, and upregulated in HOXA-AS2 knockout cells, indicating a regulatory relationship between lncRNA HOXA-AS2-miR-6867-5p-YAP1. In addition, unphosphorylated YAP is an active form, but YAP itself is a transcriptional coactivator that cannot bind DNA, so it needs to combine with other transcription factors such as TEAD1-4 to jointly initiate the transcription of downstream genes [33, 34]. Studies have shown that the downregulation of lncRNA HOXA-AS2 inhibits the phosphorylation of Yap/TAZ coactivator and its cytoplasmic retention, promotes its nuclear translocation, and promotes its function as a transcriptional coactivator, thereby promoting the expression of TEAD1-4 transcription factors. Activation. luciferase gene activity assay solution proves this. Our study also confirmed that after adding a Hippo pathway agonist, the invasive ability of gallbladder cancer cells with knockout of HOXA-AS2 were still lower. It is further proved that the Hippo signaling pathway-related regulates.
In conclusion, HOXA-AS2 may further regulate the occurrence and development of gallbladder cancer by regulating the miR-6867-5p-Yap pathway. HOXA-AS2 has important research significance in the study of potential diagnostic and therapeutic targets for gallbladder cancer. Due to the abnormal expression of HOXA-AS2 in hepatocellular carcinoma and other malignant tumors, the conclusion of this study may also be applicable to other tumor tissues. Subsequent experiments could also use the genes in this study to explore more cancer treatments. | true | true | true |
||
PMC9592237 | Xiaoshi Li | LINC01140 Targeting miR-452-5p/RGS2 Pathway to Attenuate Breast Cancer Tumorigenesis | 17-10-2022 | Background LINC01140 has been known to be involved in various cancers. However, its underlying molecular mechanism in breast cancer (BC) needs further exploration. Methods The LINC01140, miR-452-5p, and RGS2 levels in BC cells and tissues were evaluated by means of RT-qPCR and western blotting. The variations in the biological functions of BC cells were analyzed through CCK-8, transwell, western blotting, and xenograft experiments to observe cell viability, migration, levels of apoptosis-related proteins (Bax and Bcl-2), and tumor growth. The correlations existing among LINC01140, miR-452-5p, and RGS2 were validated through luciferase reporter and RIP assays. Results LINC01140 and RGS2 were remarkably downregulated in BC cells and tissues, whereas miR-452-5p was upregulated. LINC01140 overexpression diminished BC cell viability, migration, and tumor growth and facilitated apoptosis. MiR-452-5p upregulation enhanced cell viability and migration and suppressed apoptosis. Nevertheless, the additional upregulation of LINC01140 could reverse the promotive effects of miR-452-5p upregulation. Additionally, RGS2 overexpression inhibited the malignant phenotypes of BC cells, but miR-452-5p upregulation abolished this effect. In terms of mechanisms, LINC01140 acted as a miR-452-5p sponge. Moreover, RGS2 was determined to be miR-452-5p's downstream target gene in BC. Conclusion LINC01140 functioned as an antitumor agent in BC by sponging miR-452-5p to release RGS2. This hints that LINC01140 is a promising therapeutic target for BC. | LINC01140 Targeting miR-452-5p/RGS2 Pathway to Attenuate Breast Cancer Tumorigenesis
LINC01140 has been known to be involved in various cancers. However, its underlying molecular mechanism in breast cancer (BC) needs further exploration.
The LINC01140, miR-452-5p, and RGS2 levels in BC cells and tissues were evaluated by means of RT-qPCR and western blotting. The variations in the biological functions of BC cells were analyzed through CCK-8, transwell, western blotting, and xenograft experiments to observe cell viability, migration, levels of apoptosis-related proteins (Bax and Bcl-2), and tumor growth. The correlations existing among LINC01140, miR-452-5p, and RGS2 were validated through luciferase reporter and RIP assays.
LINC01140 and RGS2 were remarkably downregulated in BC cells and tissues, whereas miR-452-5p was upregulated. LINC01140 overexpression diminished BC cell viability, migration, and tumor growth and facilitated apoptosis. MiR-452-5p upregulation enhanced cell viability and migration and suppressed apoptosis. Nevertheless, the additional upregulation of LINC01140 could reverse the promotive effects of miR-452-5p upregulation. Additionally, RGS2 overexpression inhibited the malignant phenotypes of BC cells, but miR-452-5p upregulation abolished this effect. In terms of mechanisms, LINC01140 acted as a miR-452-5p sponge. Moreover, RGS2 was determined to be miR-452-5p's downstream target gene in BC.
LINC01140 functioned as an antitumor agent in BC by sponging miR-452-5p to release RGS2. This hints that LINC01140 is a promising therapeutic target for BC.
Breast cancer (BC) is among the most prevalent cancers among women around the world [1, 2]. It accounts for 18% of all cancer cases in women with approximately 1 million new cases reported worldwide each year [1, 2]. Over the past few decades, though, breast cancer mortality has declined steadily, thanks to improved screening and treatment [3]. But because this type of cancer can spread to distant organs or lymph nodes, it is still considered the leading cause of death in women [4, 5]. Traditional therapies, including chemotherapy, endocrine therapy, and radiotherapy, do not cure BC most of the time but only improve clinical outcomes [6]. Hence, it is urgent to develop effective approaches for BC diagnosis and therapy. Long noncoding RNAs (lncRNAs) are larger than 200 bp and have no protein-coding ability [7]. Previous studies have found that abnormal lncRNA expression often has a specific pattern in human tumor tissues [8]. Over the recent years, there has been accumulating evidence that lncRNA is new cancer mediator that can regulate gene expression through transcription, posttranscriptional, or epigenetic levels, in order to participate in almost all malignant behaviors of tumor cells [8, 9]. For example, the newly discovered LINC02273 is significantly elevated in metastatic BC lesions and is an independent prognostic factor in predicting patient survival [10]. LncRNA HOTAIR induces BC cell survival and optimizes drug resistance [11]. LINC01140 is a potential regulator of cancer drive or containment in different tumors. LINC01140 inhibits the progression of lung cancer [12] and sarcomas [13], whereas it promotes bladder cancer. Li et al. have analyzed the cBioPortal database and reported that LINC01140 expression is significantly reduced in the tumor samples from BC patients [14]. This predicted a worse recurrence-free survival rate of BC patients [14]. Nevertheless, the underlying mechanisms of LINC01140 need to be researched further. microRNAs (miRNAs) are small RNA molecules that bind to targets of mRNA, acting as gene silencing and translation inhibitors [15]. Over the past few years, the function of miRNAs in biological processes and in the occurrence and development of various human diseases, including cancer, has been extensively studied [16, 17]. MiR-452-5p has been proven to be a regulatory factor of colorectal cancer, liver cancer, and lung cancer, specifically promoting the invasion, migration, and proliferation of cancer cells [18, 19]. Furthermore, the elevated miR-452-5p expression in squamous cell carcinoma has been reported to have a significant involvement in tumor lymph node metastasis [20]. More importantly, one analysis has indicated that miR-452-5p is aberrantly expressed in BC [21]. Hence, I speculate that miR-452-5p may play an important role in BC. I expect to further explore miR-452-5p's influence on BC development and its specific regulatory mechanism. Regulator of G-protein signaling 2 (RGS2) is a member of the GTPase activating protein (GAP) family of Ga subunits. It has been originally identified as an inhibitor of G protein signal transduction, but recent studies have shown that it has cellular proliferation regulatory functions as well [22]. RGS2 is highly expressed in normal human cells but downregulated in cancer cells, including breast adenocarcinoma, wherein it plays a tumor suppressive function [23, 24]. Therefore, in-depth study of the effects of RGS2 may improve my knowledge of the pathogenesis of BC. The potential mechanism of LINC01140 in BC needs to be analyzed in detail. I hypothesize that the LINC01140/miR-452-5p/RGS2 axis is a new signaling pathway related to the progression of BC. This may provide valuable theoretical basis for BC diagnosis and treatment.
Thirty-eight [25] sets of tumoral and normal adjacent tissues were acquired from BC patients. All patients that received surgery in the First Affiliated Hospital of Chengdu Medical College provided a signed informed consent and did not undergo radiotherapy or chemotherapy prior their procedure. The cancer tissues were confirmed by at least two pathologists. The Ethics Committee of the First Affiliated Hospital of Chengdu Medical College approved this study.
Three of BC cell lines (MDA-MB-231, MCF-7, and HCC1937) and MCF-10A were all obtained from ATCC (USA). DMEM (Gibco, USA) containing 1% P/S as well as 10% FBS (Gibco) was utilized for the culturing of cells. The cultures were maintained at 37°C in humidified atmosphere that contained 5% CO2. Cells were digested and subcultured when the confluence reached 80-90%. The third-generation cells at logarithmic growth phase were chosen for the succeeding studies.
MiRNA was isolated using the NucleoSpin® miRNA kit (Macherey Nagel, France), and total RNA was extracted with the aid of the RNA isolation kit (Takara, Japan). ImProm-II reverse transcription system (Promega, USA) and PrimeScript™ RT master mix (Takara) were utilized for miRNA and RNA reverse transcription, respectively. Subsequently, LightCycler 480 System (Roche, Germany) with the miRcute miRNA qPCR detection kit (Tiangen) or TaqMan universal master mix (Applied Biosystems) was used for the miRNA and RNA qPCR assay in a CFX connect real-time PCR detection system (Bio-Rad, USA). The RNA and miRNA levels were computed by applying the 2-ΔΔCt formula, and the results were normalized against GAPDH and U6, respectively [26]. Table 1 lists the primers used.
Following the included protocol, the PARIS kit (Thermo Fisher Scientific, USA) was employed to isolate and collect the nucleus and cytoplasm of MDA-MB-231 and HCC1937 cells. The LINC01140 expressions in the nucleus and cytoplasm were assessed via RT-qPCR and then normalized against U6 and GAPDH, respectively.
LINC01140 and RGS2 overexpression plasmids (LINC01140-OE and RGS2-OE), as well as the empty vector, were provided by GeneChem (Shanghai, China). The plasmids were utilized for the overexpression of LINC01140 and RGS2, and the empty vector served as the control. SwitchGear Genomics (USA) offered the miR-452-5p mimic and its negative control, mimic-NC. Lipofectamine 2000 (Invitrogen, USA) was utilized in transfecting the MDA-MB-231 and HCC1937 cells with 100 nM mimic or 2 μg/ml OE RNA. After even and careful shaking, the culture was incubated at 37°C for 48 h. Afterward, RT-qPCR assay was performed to evaluate the transfection efficiency.
CCK-8 kit (Sigma, USA) was utilized to observe the viabilities of the HCC1937 and MDA-MB-231 cells. Simply put, 5 × 104 cells were plated on each of the 96-wells of a culture plate and then cultivated for 0, 24, 48, and 72 h. Afterward, CCK-8 (10 μl) was added and incubated with the transfected cells at 37°C for 2 h. The wavelengths were determined using the FLx800 fluorescence microplate reader (BioTek, USA) with a 450-nm filter.
Transwell inserts (Corning, USA) were utilized in performing this assay. Complete medium (500 μl) was poured into the lower compartments of the inserts. Meanwhile, 1 × 105 treated cells/ml were mixed with a medium that was serum-free and then pipetted into the upper chambers. After 24 h of culturing, the migrated cells were immobilized using 4% methanol for 15 min and then dyed with 0.25% crystal violet for 10 min. The cells that migrated were photographed and tallied with the help of a microscope (Olympus, Japan).
Proteins were isolated from the MDA-MB-231 and HCC1937 cell lines with the aid of a RIPA lysis buffer (Invitrogen). Their concentrations were then quantified with a BCA kit (Thermo Fisher). The protein samples were subjected to electrophoresis on a 10% SDS-PAGE gel before they were moved onto PVDF membranes. They then were blocked with 5% nonfat milk for 1 h at 25°C. Subsequently, the proteins were incubated overnight at 4°C with the following antibodies from Abcam: anti-Bax (1 : 1000, ab32503), anti-Bcl-2 (1 : 1000, ab32124), anti-RGS2 (1 : 1000, ab155762), and anti-GAPDH (1 : 2000, ab181603). Afterward, they were supplemented with the corresponding secondary antibody (1 : 2000, ab97051, Abcam) and maintained at room temperature for 2 hours. At last, the protein blots were visualized with the aid of an ECL Reagent (GE Healthcare, USA).
HCC1937 cells (1 × 106) stably transfected with LINC01140-OE or empty vector were subcutaneously injected into ten BALB/c mice (5 weeks old, five mice/group) acquired from Hunan SJA Laboratory Animal (China). The tumors were measured every 4 days, and their volumes were calculated using the following formula: volume = (length) × (width)2/2. The mice were sacrificed 28 days after the inoculation of tumor cells, and the tumor xenografts were excised then weighed. The Animal Care and Use Committee of the First Affiliated Hospital of Chengdu Medical College has authorized this assay.
Wild-type LINC01140 (LINC01140-WT) or RGS2 3'UTR (RGS2-WT) sequences with miR-452-5p binding sites were inserted into pGL3 reporter constructs (Promega). Mutant LINC01140 (LINC01140-MUT) or RGS2 (RGS2-MUT) were obtained with the QuikChange site-directed mutagenesis kit (Stratagene, USA). Lipofectamine 2000 was then utilized to deliver the reporter constructs, along with a mimic-NC or miR-452-5p mimic, into the HCC1937 and MDA-MB-231 cells. Forty-eight hours later, the luciferase activities of the different vectors were revealed by the dual-luciferase report analysis system (Promega, USA).
This assay was conducted using the BersinBioTM RIP kit (BersinBio, China). Magnetic beads were pretreated with Ago2 and negative control IgG antibody. Afterward, they were conjugated with prepared HCC1937 and MDA-MB-231 cell lysate suspension for 4 h at 4°C. Protein A-Sepharose was then added to the product and then maintained at 4°C for 4 h. Finally, RT-qPCR was conducted to assess LINC01140 and miR-452-5p levels.
The experimental data were expressed as the mean ± SD and were analyzed in GraphPad Prism 8.0 (GraphPad Software, USA). Correlation analyses between miR-452-5p and LINC01140 or RGS2 levels in cancer tissues were determined by Pearson's correlation coefficient. P < 0.05 indicated statistical significance. Student's t-test was used to test the variations between two groups, whereas one-way ANOVA was applied for multiple groups.
As illustrated in Figure 1(a), the expressions of LINC01140 in MCF-7, MDA-MB-231, and HCC1937 cells are 25%, 55%, and 75% lower, respectively, than that of the MCF-10A cells. I also determined LINC01140 levels in clinical tissues, and the data manifested that the LINC01140 levels in BC specimens were approximately 80% lower than that of the normal tissues (Figure 1(b)). Subcellular localization experiment was conducted to identify the location of LINC01140 in BC cells. The outcome of the experiment revealed that LINC01140 predominantly existed more within the cytoplasm than in the nucleus. This suggests that LINC01140 may perform posttranscriptional and transcriptional regulatory functions in BC cells (Figure 1(c)). Based on these results, I clarified LINC01140's influence on the behavior of BC cells. LINC01140 overexpression vector was delivered into the HCC1937 and MDA-MB-231 cells. Thereafter, LINC01140 levels increased by more than 7-fold in the BC cell lines in contrast to that in the empty vector group (Figure 1(d)). CCK-8 analysis revealed that the overexpression of LINC01140 reduced cell viability by about 40% (Figure 1(e)). Moreover, the transwell experiment demonstrated that LINC01140 upregulation reduced the number of migrating cells by over 45% (Figure 1(f)). In addition, I also observed an increase in Bax protein levels and a decrease of Bcl-2 in the LINC01140-OE groups (Figure 1(g)). These abovementioned results manifest that LINC01140 overexpression effectively represses BC cell survival in vitro. I further explored LINC01140's impact on the growth of BC cells in vivo. HCC1937 cells from the LINC01140-OE or empty vector groups were administered into the nude mice. In comparison with empty vector groups, the tumor sizes and volumes among the nude mice from the LINC01140-OE group were reduced (Figures 2(a) and 2(b)). Weighing the tumors further revealed that LINC01140 upregulation diminished the weight of the tumors (Figure 2(c)). These suggest that the ectopic expression of LINC01140 may inhibit the growth of BC cells in vivo.
In virtue of starBase, the miRNAs binding to LINC01140 were predicted (Supplementary table 1). Due to the inhibitory effect of miR-452-5p in multiple cancers, I selected miR-452-5p to explore its regulatory mechanism in breast cancer. The binding site between LINC01140 and miR-452-5p sequences is shown in Figure 3(a). Next, I explored the binding relationship between LINC01140 and miR-452-5p. Results of the luciferase assay showed that the miR-452-5p ectopic expression only reduced LINC01140-WT luciferase activity, whereas that of LINC01140-MUT did not change (Figure 3(b)). Additionally, it was uncovered via RIP experiment that miR-452-5p and LINC01140 were enriched in compounds precipitated by anti-Ago2 antibodies (Figure 3(c)). This implies that miR-452-5p binds to LINC01140. Hence, I assumed that miR-452-5p was abnormally expressed in BC. It was shown via RT-qPCR that miR-452-5p levels in BC were approximately 4.6-fold of that of the normal samples (Figure 3(d)). Pearson correlation analysis then confirmed an inverse association between LINC01140 and miR-452-5p levels in BC samples (Figure 3(e)). Additionally, I also observed significantly elevated miR-452-5p levels among the BC cells lines compared to MCF-10A cells (Figure 3(f)). Moreover, I studied the regulatory effect of LINC01140 on miR-452-5p and learned that the expression of miR-452-5p was downregulated by more than 80% after LINC01140-OE transfection. Furthermore, the miR-452-5p upregulation effected by the miR-452-5p mimic, which was over 7.5-fold of the control, was also reversible by LINC01140-OE (Figure 3(g)). In a summary, miR-452-5p was LINC01140's target gene.
In a follow-up study, I used a salvage trial to verify the role of LINC01140/miR-452-5p in BC. First, as shown in Figure 4(a), CCK-8 reveals that miR-452-5p overexpression increased cell viability by about 1.4 times, but the extra LINC01140 upregulation offsets this effect on cell viability. In addition, miR-452-5p mimic inhibited Bax protein levels and upregulated Bcl-2 levels, but this was reversed by the enhanced apoptotic effect of LINC01140-OE (Figure 4(b)). Moreover, in transwell analysis, cell migration levels were found to be upregulated by at least 2-fold after miR-452-5p mimic transfection. However, LINC01140 overexpression restored this upregulation to normal levels (Figure 4(c)). In summary, I conclude that LINC01140 modulates miR-452-5p to achieve its anticancer function in BC cells.
To explore the downstream of miR-452-5p, I utilized starBase to forecast its target genes. I discovered that RGS2 has a binding site for miR-452-5p (Figure 5(a)). Targeting analyses revealed that RGS2-WT group decreased the luciferase activity by 55% after the transfection of the miR-452-5p mimic. Meanwhile, no significant changes in luciferase activity were observed in the RGS2-MUT group (Figure 5(b)). Moreover, the expression of RGS2 mRNA in BC tissues was 45% of that in normal tissues (Figure 5(c)). I further found that HCC1937 and MDA-MB-231 cells manifested RGS2 levels that were 60% and 70% lower, respectively, than that in the MCF-10A cells (Figure 5(d)). Also, miR-452-5p levels were inversely correlated with RGS2 levels (Figure 5(e)). Western blotting showed that miR-452-5p mimic treatment downregulated RGS2 protein levels. Nevertheless, this change could be reversed by RGS2 overexpression (Figure 5(f)). These results suggest that miR-452-5p targets and negatively regulates RGS2.
I scrutinized the miR-452-5p and RGS2's interaction in malignant BC phenotype. CCK-8 demonstrated that in contrast to the empty vector group, cell viability in the RGS2-OE group was lower, but it was also partially reversible through the introduction of the miR-452-5p mimic (Figure 6(a)). In addition, after RGS2 upregulation, Bax protein levels increased while that of Bcl-2 decreased. However, miR-452-5p mimic could considerably mitigate the proapoptosis effect engendered by RGS2 upregulation (Figure 6(b)). The transwell experiments showed that the cell migration level diminished after RGS2-OE transfection. Meanwhile, miR-452-5p mimic alleviated the repressive influence of RGS2 upregulation on cell migration (Figure 6(c)). These reveal that RGS2 restrains the aggressive behaviors of BC cells but is mediated by miR-452-5p.
BC has killed about 630,000 people worldwide, according to Globocan data in 2018 [27]. Despite significant advances in cancer medicine over the past decade, my understanding of BC progression remains limited [28]. It would be helpful to explore the genetic characteristics of BC and its relationship with tumor progression to guide the clinical treatment and diagnosis of BC [29]. In my research, I observed that LINC01140 was remarkably downregulated in BC cells and tissues. Moreover, its overexpression inhibited BC cell viability, migration, and tumor growth, while promoting apoptosis. LINC01140 may be a tumor suppressor associated with BC. In terms of mechanisms, LINC01140 acts as miR-452-5p'sponge, releasing RGS2. These results may contribute theoretical bases for the clinical diagnosis and treatment of BC. LINC01140 is one of the few new lncRNAs identified to be involved in tumor prognosis. Hu et al. [13] have discovered that the low levels of LINC01140 in metastatic sarcomas indicate disease-free survival, disease-specific survival, and poor overall survival. Meanwhile, its high expression may promote survival of sarcomas. Inconsistently, low LINC01140 inhibits the survival and metastasis of glioma cells. A study by Wu et al. [30] revealed that LINC01140 was significantly upregulated in muscle-invasive bladder cancer. Furthermore, its downregulation inhibited the invasive ability of cancer cells [30]. This may be caused by the tissue specificity of LINC01140, suggesting that LINC01140 may operate as a tumor suppressor or progenitor in different tumor types. Herein, my experimental results confirmed the previous report by Li et al. [14] that LINC01140 expression was significantly reduced in BC. Furthermore, in the present study, I found that the overexpression of LINC01140 inhibited the malignant behavior of BC cells in vitro and reduced tumor growth in nude mice xenografts. This reflects that LINC01140 may serve as a potential regulator gene for the suppression of BC. A growing number of researchers are paying attention to the role of miRNAs in human diseases and cancer [31]. Some works have documented that miRNA dysregulation has a crucial part in the progression of BC [32]. It can be used as a diagnostic marker for BC and is related to the treatment resistance of BC [32]. For example, the miR-205 expressions among the different subtypes of BC, ranging from the less aggressive subtype to the more aggressive triple negative breast cancer (TNBC), affect metastatic ability, treatment response, and patient survival [33]. MiR-9 levels may be used as a potential noninvasive tumor marker for BC [34]. Li et al. [35] reported that the results of the next-generation sequencing showed the overexpression of miR-452 BC tissues. Meanwhile, their RT-qPCR results revealed the low levels of miR-452 levels in TNBC. The researchers also reported the inhibitory influence of miR-452 antagonists on TNBC xenografts. Herein, my findings show high expressions of miR-452-5p in BC, which is similar to the next-generation sequencing results of Li et al. [35], but contrary to the results of RT-qPCR detection. Nevertheless, it is speculated that the difference in the aggressiveness of the cancer tissue may be responsible for this. MiR-452-5p is an oncogenic factor in colorectal cancer, liver cancer, and lung cancer. It promotes cell proliferation, cell cycle transformation, and chemotherapy resistance, but inhibits apoptosis [18–20]. In contrast, however, miR-452-5p inhibits the development of prostate and cervical cancers [36, 37]. These may be due to the heterogeneity of the cancers. The outcomes of this study are consistent with those in colorectal cancer, liver cancer, and lung cancer. I have learned that the miR-452-5p upregulation promotes the tumorigenicity of BC cells, thus revealing that miR-452-5p facilitates the progression of BC. The lncRNA-miRNA-mRNA network has been extensively studied as biomarkers and potential therapeutic targets for BC diagnosis and prognosis [25, 38]. For instance, BCRT1 promotes BC progression by sponging miR-1303 and upregulating PTBP3 [39]. OIP5-AS1 and miR-216a-5p/GLO1 form a competitive endogenous RNA (ceRNA) regulatory network in BC, thus promoting the malignant behavior of cancer cells [40]. In my current work, I further explored miR-452-5p's interaction with LINC01140. Targeting analyses uncovered the miR-452-5p's binding to LINC01140. More interestingly, I learned that miR-452-5p levels were negatively associated to that of LINC01140 in BC tissues. This further supports that LINC01140 has a sponging effect on miR-452-5p. Salvage analyses exhibited that overexpressing LINC01140 counterbalanced the malignant migration and proliferation of BC cell lines after the increase of endogenous miR-452-5p. Therefore, I hypothesized that LINC01140 could inactivate miR-452-5p by sponging it to inhibit the induction of malignant BC cell behavior. The abnormal regulation of RGS2 has been associated with the development of solid tumors. Also, the downregulation of RGS2 has been reported in the progression of various cancers [23, 24, 41]. RGS2 is underexpressed in BC; its overexpression can inhibit epidermal growth factors or the serum-induced proliferation of cancer cells [24]. BC patients with poor levels of RGS2 had notably lower overall survival as RGS2 operates as a suppressor gene in BC [42]. In my current work, I confirmed the low expression of RGS2 in BC. I also revealed that RGS2 overexpression inhibited the malignant behavior of BC cells, and that it had an anticancer effect in BC. Moreover, I evidenced that RGS2 was miR-452-5p's target gene and was involved in the proliferation and migration of BC cells, which were regulable via the LINC01140/miR-452-5p axis. Given the limitations of my study, it is clear that further research needs to be done. Cumulative studies have documented abnormal miRNA cluster expressions in BC, showing protumor and antitumor effects [32]. In future studies, I may explore the effects of miR-452 clusters on BC, based on the report of Li et al. [35]. Furthermore, the downstream signaling pathway of the LINC01140/miR-452-5p/RGS2 axis is also worth investigating. In conclusion, I emphasize the important role of LINC01140 in the emergence and progression of BC. My study is the first to demonstrate that LINC01140 is involved in BC migration and survival through the miR-452-5p/RGS2 axis. The findings from this study contribute new insights on the mechanisms of BC progression and suggest potential targets for its treatment. | true | true | true |
||
PMC9592412 | Qiang Shu,Xiaoling Liu,Jushu Yang,Tinggang Mou,Fei Xie | The clinical prognostic value of lncRNA FOXP4-AS1 in cancer patients: A meta-analysis and bioinformatics analysis based on TCGA datasets | 21-10-2022 | cancer,disease free survival,FOXP4-AS1,long non-coding RNA,overall survival,prognosis | Background: The mortality and recurrence of patients with cancer is of high prevalence. Long non-coding RNA (lncRNA) forkhead box P4 antisense RNA 1 (FOXP4-AS1) is a promising lncRNA. There is increasing evidence that lncRNA FOXP4-AS1 is abnormally expressed in various tumors and is associated with cancer prognosis. This study was designed to identify the prognostic value of lncRNA FOXP4-AS1 in human malignancies. Methods: We searched electronic databases up to April 29, 2022, including PubMed, Cochrane Library, Embase, MEDLINE, and Web of Science. Eligible studies that evaluated the clinicopathological and prognostic role of lncRNA FOXP4-AS1 in patients with malignant tumors were included. The pooled odds ratios (ORs) and the hazard ratios (HRs) were calculated to assess the role of lncRNA FOXP4-AS1 using Stata/SE 16.1 software. Results: A total of 6 studies on cancer patients were included in the present meta-analysis. The combined results revealed that high expression of lncRNA FOXP4-AS1 was significantly associated with unfavorable overall survival (OS) (HR = 1.99, 95% confidence interval [CI]: 1.65–2.39, P < .00001), and poor disease-free survival (DFS) (HR = 1.81, 95% CI: 1.54–2.13, P < .00001) in a variety of cancers. In additional, the increase in lncRNA FOXP4-AS1 expression was also correlated with tumor size ((larger vs smaller) (OR = 3.16, 95% CI: 2.12–4.71, P < .00001), alpha-fetoprotein (≥400 vs <400) (OR = 3.81, 95%CI: 2.38–6.11, P = .83), lymph node metastasis (positive vs negative) (OR = 2.93, 95%CI: 1.51–5.68, P = .001), and age (younger vs older) (OR = 2.06, 95% CI: 1.41–3.00, P = .00002) in patients with cancer. Furthermore, analysis results using The Cancer Genome Atlas (TCGA) dataset showed that the expression level of lncRNA FOXP4-AS1 was higher in most tumor tissues than in the corresponding normal tissues, which predicted a worse prognosis. Conclusions: In this meta-analysis, we demonstrate that high lncRNA FOXP4-AS1 expression may become a potential marker to predict cancer prognosis. | The clinical prognostic value of lncRNA FOXP4-AS1 in cancer patients: A meta-analysis and bioinformatics analysis based on TCGA datasets
The mortality and recurrence of patients with cancer is of high prevalence. Long non-coding RNA (lncRNA) forkhead box P4 antisense RNA 1 (FOXP4-AS1) is a promising lncRNA. There is increasing evidence that lncRNA FOXP4-AS1 is abnormally expressed in various tumors and is associated with cancer prognosis. This study was designed to identify the prognostic value of lncRNA FOXP4-AS1 in human malignancies.
We searched electronic databases up to April 29, 2022, including PubMed, Cochrane Library, Embase, MEDLINE, and Web of Science. Eligible studies that evaluated the clinicopathological and prognostic role of lncRNA FOXP4-AS1 in patients with malignant tumors were included. The pooled odds ratios (ORs) and the hazard ratios (HRs) were calculated to assess the role of lncRNA FOXP4-AS1 using Stata/SE 16.1 software.
A total of 6 studies on cancer patients were included in the present meta-analysis. The combined results revealed that high expression of lncRNA FOXP4-AS1 was significantly associated with unfavorable overall survival (OS) (HR = 1.99, 95% confidence interval [CI]: 1.65–2.39, P < .00001), and poor disease-free survival (DFS) (HR = 1.81, 95% CI: 1.54–2.13, P < .00001) in a variety of cancers. In additional, the increase in lncRNA FOXP4-AS1 expression was also correlated with tumor size ((larger vs smaller) (OR = 3.16, 95% CI: 2.12–4.71, P < .00001), alpha-fetoprotein (≥400 vs <400) (OR = 3.81, 95%CI: 2.38–6.11, P = .83), lymph node metastasis (positive vs negative) (OR = 2.93, 95%CI: 1.51–5.68, P = .001), and age (younger vs older) (OR = 2.06, 95% CI: 1.41–3.00, P = .00002) in patients with cancer. Furthermore, analysis results using The Cancer Genome Atlas (TCGA) dataset showed that the expression level of lncRNA FOXP4-AS1 was higher in most tumor tissues than in the corresponding normal tissues, which predicted a worse prognosis.
In this meta-analysis, we demonstrate that high lncRNA FOXP4-AS1 expression may become a potential marker to predict cancer prognosis.
Cancer is one of the leading causes of death worldwide.[ However, the exact mechanism underlying this cancer remains unclear. Furthermore, surveillance of patients with early stage cancer remains difficult. Hence, many cancer cases are identified at an advanced stage and have a dismal prognosis. The official databases of the World Health Organization and American Cancer Society indicate that cancer poses the highest clinical, social, and economic burden among all human diseases. A total number of 18 million new cases have been diagnosed in 2018, the most frequent of which are lung (2.09 million cases), breast (2.09 million cases), and prostate (1.28 million cases) cancers.[ Therefore, early diagnosis and intervention have become vital for improving the overall survival (OS) of patients with cancer. Long non-coding RNA (lncRNA) are defined as transcripts >200 nucleotides in length that display limited protein-coding potential.[ In recent years, lncRNAs have been found to play significant regulatory roles in a variety of diseases, especially in the biological processes of malignant tumors, including differentiation, migration, apoptosis, and dose compensation effects.[ Recent studies have proposed that LINC00675 is related to clinicopathological features and prognosis of various cancer patients by participating in cancer cell proliferation and invasion.[ In cervical cancer, SIP1 expression is upregulated by lncRNA NORAD to promote proliferation and invasion of cervical cancer cells.[ Studies have shown that lncNONHSAAT081507.1 (LINC81507) plays an inhibitory role in the progression of non-small cell lung cancer and acts as a therapeutic target and potential biomarker for the diagnosis and prognosis of nonsmall cell lung cancer.[ These results provide evidence that lncRNAs may serve as novel prognostic biomarkers and therapeutic targets in human tumors.[ Forkhead box P4 antisense RNA 1 (FOXP4-AS1), a member of the lncRNA family, is an antisense lncRNA to FOXP4. The lncRNA FOXP4-AS1, which is an lncRNA related to tumors, is believed to participate in the occurrence of tumors and promote tumor proliferation, invasion, and migration. Extensive studies have indicated that FOXP4-AS1 is highly expressed in several malignancies, including hepatocellular carcinoma (HCC),[ colorectal carcinoma,[ and nasopharyngeal carcinoma (NPC).[ Thus, its upregulation is usually related to tumor grade and poor prognosis.[ However, no systematic meta-analysis has yet been conducted to support the prognostic value of FOXP4-AS1 in these cancers. Hence, a meta-analysis was performed to investigate the clinical prognostic value of lncRNA FOXP4-AS1 in patients with cancer.
This meta-analysis was conducted in accordance with the Guidelines for Preferred Reports of Systematic Reviews and Meta-Analysis and Meta-analysis of Observational Epidemiological Studies statements. A comprehensive literature search was conducted. In order to identify relevant articles, 2 reviewers independently searched electronic databases, including PubMed, Cochrane Library, EMBASE, Medline and Web of Science.Use the following search terms: “long non-coding RNA FOXP4-AS1,” “lncRNA FOXP4-AS1,” “Forkhead box P4 antisense RNA 1,” “tumor,” “cancer,” and “prognosis.” The issue will be discussed with a third reviewer if there are any inconsistencies. To select eligible studies, we used the following criteria: definitive diagnosis or histopathological diagnosis in cancer patients; survival and clinical prognostic parameters of lncRNA FOXP4-AS1 in cancer patients were reported. The combined disaster risk (HR) and 95% confidence intervals (CI) were calculated using sufficient information. The exclusion criteria were as follows: studies without prognostic outcome information, non-human studies, letters, case reports, review articles, duplicate publications, and studies without original data.
The following information was extracted from each study by 3 independent authors and a consensus was reached: author, country, publication year, tumor type, cancer size, follow-up time, detection method, and cutoff value. Based on distant metastasis (DM), tumor size, and cancer node metastasis stage, the number of patients was divided for each group, and the number of patients with high or low FOXP4-AS1 expression in each group. When the HR with 95% CI was reported in a univariate or multivariate analysis, it was directly extracted from the report (https://sourceforge.net/projects/digitizer/).[ We used Engauge Digitizer to calculate HR and 95% CI based on Kaplan–Meier survival curves, and quality assessments for all included studies were based on the Newcastle–Ottawa quality assessment scale, which includes 3 dimensions: selection, comparability, and exposure. Each study was given a score ranging from 0 to 9. A study with a Newcastle–Ottawa quality assessment scale score of ≥6 was considered to be of high quality.[
All statistical analyses of the data were performed using Review Manager (RevMan) 5.3 software and Stata/SE 16.1 software. Sensitivity analysis was performed by omitting the literature one by one to determine whether the results were stable, and the publication bias of this meta-analysis was evaluated using the Begg test according to Stata/SE 16.1 software. The Q test and I2 statistics were applied to estimate the heterogeneity of the results. A fixed-effects model was selected when an I2 < 50% was observed. The synthetic estimate was calculated based on the random-effects model when heterogeneity was evident (I2 > 50%). Statistical significance was set at P < .05.
After the preliminary online search, the investigators retrieved 110 relevant studies from electronic databases. After the removal of duplicates, 57 studies were excluded. After thorough screening of titles and abstracts, 14 publications were included. After carefully assessing the full texts, 6 studies published between 2018 and 2021 were included in the present meta-analysis. The literature screening process is shown in Figure 1. These eligible studies included 1128 patients. In the present meta-analysis, a variety of tumor types were reported, including HCC,[ NPC,[ mantle cell lymphoma (MCL),[ colorectal cancer,[ and osteosarcoma.[ The expression of lncRNA FOXP4-AS1 in these included studies was quantified using real-time fluorescent PCR. The median was selected as the cutoff value to distinguish between high and low expression of lncRNA FOXP4-AS1. Six eligible studies used the OS to estimate patient survival. The detailed clinical characteristics of the patients are summarized in Table 1.
Overall, all the included studies investigated cancer prognosis. A total of 1128 patients were assessed for HR and 95% CI for OS. A random-effects model was used to analyze the pooled HR, and its 95% CI depended on obvious heterogeneity (P = .06, I2 = 53%). We further elucidated the relationship between FOXP4-AS1 expression and OS, as illustrated in Figure 2. The pooled results revealed that high expression of lncRNA FOXP4-AS1 was related to poor prognosis of cancers (HR = 1.99, 95% CI:1.65–2.39, P < .00001, Fig. 2A). In terms of DFS, 5 studies were included, and the pooled results indicated that patients with high expression of lncRNA FOXP4-AS1 had poor DFS (HR = 1.81, 95% CI:1.54–2.13, P < .00001, Fig. 2B). Thus, the prognosis of cancer patients with lncRNA FOXP4-AS1 overexpression was worse than that of patients with low lncRNA FOXP4-AS1 expression. These results indicate that lncRNA FOXP4-AS1 may be a factor in predicting the prognosis of cancer patients.
According to the evaluation of the 6 eligible studies that contained detailed clinicopathological characteristics, it was observed that the elevated expression of lncRNA FOXP4-AS1 positively correlated with tumor size (larger vs smaller) (OR = 3.16, 95%CI: 2.12–4.71, P < .00001, Fig. 3B). In particular, the overexpression of lncRNA FOXP4-AS1 was consistent with elevated alpha-fetoprotein (OR = 3.81, 95%CI: 2.38–6.11, P = .83, Fig. 3C) in HCC, and the fixed effects model was selected to estimate due to the inconspicuous heterogeneity. The analysis results revealed that the patients with lncRNA FOXP4-AS1 overexpression were more vulnerable to younger (OR = 2.06, 95%CI: 1.41–3.00, P = .00002, Fig. 3A) and lymph node metastasis (OR = 2.93, 95%CI: 1.51–5.68, P = .001, Fig. 3D) in patients with cancer. However, there were no significant differences between lncRNA FOXP4-AS1 expression and TNM stage (OR = 1.38, 95%CI: 0.42–4.48, P = .59, Fig. 4A), DM (OR = 0.84, 95%CI: 0.49–1.45, P = .54, Fig. 4B), gender (OR = 1.08, 95%CI: 0.70–1.67, P = .72, Fig. 4C), differentiation (OR = 0.91, 95%CI: 0.49–1.67, P = .76, Fig. 4D). This information is presented in Table 2.
This meta-analysis evaluated the publication bias using the Begg test. The results showed no significant publication bias affecting OS analysis (Pr > IzI = 0.368) (Fig. 5). Figure 6 illustrates the sensitivity analysis we conducted to show that the HRs were robust even after removing all the studies individually.
As well, the investigators made use of TCGA dataset to analyze the expression of lncRNA FOXP4-AS1 in the different types of cancers. This database consists of 22 different types of human malignant tumors, including bladder urothelial carcinoma, breast invasive carcinoma, cervical squamous cell carcinoma and endocervical adenocarcinoma, cholangiocarcinoma, colon adenocarcinoma, esophageal carcinoma (ESCA), head and neck squamous cell carcinoma (HNSC), kidney chromophobe (KICH), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), pancreatic adenocarcinoma (PAAD), prostate adenocarcinoma (PRAD), rectum adenocarcinoma (READ), sarcoma (SARC), skin cutaneous melanoma (SKCM), stomach adenocarcinoma (STAD), thyroid carcinoma (THCA), thymoma (THYM), and uterine corpus endometrial carcinoma (UCEC). As shown in Figure 7, lncRNA FOXP4-AS1 was significantly overexpressed in tumor tissues, especially in patients with COAD, PRAD, READ, and STAD. We found that the expression of lncRNA FOXP4-AS1 in 6 malignant tumors was significantly different in clinical staging, such as KIRP, KIPAN (pan-kidney cohort), HNSC, KIRC, LIHC, and TGCT (testicular germ cell tumors) (Fig. 8). Moreover, the expression of lncRNA FOXP4-AS1 was significantly different in the differentiation of 6 malignant tumors, including ESCA, stomach and esophageal carcinoma (STES), STAD, HNSC, PAAD, and ovarian serous cystadenocarcinoma (OV) (Fig. 9). Additionally, the investigators explored whether lncRNA FOXP4-AS1 expression was associated with the survival and prognosis of cancer patients in the TCGA dataset. The results revealed that upregulated lncRNA FOXP4-AS1 expression in different types of malignant tumors exhibited negative effects on OS (Fig. 10) and DFS (Fig. 11).
As is well known, numerous cancers have a poor prognosis because early diagnosis is difficult. A high proportion of patients have an advanced stage of cancer at diagnosis, indicating that tumors have spread to adjacent or distant organs, tissues, and lymph nodes, indicating poor prognosis. Consequently, it is indispensable to develop novel biomarkers that are reliable for predicting cancer patient diagnosis and prognosis. In recent years, human cancers can be predicted with the help of lncRNAs. The lncRNA FOXP4-AS1, in particular, has generated a lot of interest because accumulating studies suggest that it may play a key role in determining the clinical outcome of different types of cancers.[ In most studies, there are a limited number of samples; therefore, more evidence is needed regarding the prognostic role of lncRNA FOXP4-AS1, which provides sufficient data for further investigation. As far as we are aware, this is the first meta-analysis that provides insights regarding the precise role played by lncRNA FOXP4-AS1 in patient survival and clinicopathological parameters. The present study demonstrated that elevated lncRNA FOXP4-AS1 levels were significantly associated with inferior OS and DFS in various cancers, indicating that lncRNA FOXP4-AS1 may serve as an indicator for cancer prognosis, with the potential to support new therapies. On the basis of clinicopathological features, patients with high lncRNA FOXP4-AS1expression are more inclined to have a high risk of tumor growth than those with low lncRNA FOXP4-AS1 expression. Additionally, the inferiority of high lncRNA FOXP4-AS1 expression on alpha-fetoprotein and lymph node metastasis was observed, indicating that lncRNA FOXP4-AS1 overexpression was associated with worse clinicopathological characteristics. However, the high expression of lncRNA FOXP4-AS1 was not associated with gender, DM, differentiation, or tumor node metastasis stage. In cancers, abnormal expression of the lncRNA FOXP4-AS1 is associated with poor clinical prognosis, and the exact underlying mechanisms still need to be clarified. For accounting this significant association, there are several possible explanations. Wang et al proposed that lncRNA FOXP4-AS1 is involved in HCC development by mediating in the PI3K-Akt signaling Pathway.[ Tao et al demonstrated that lncRNA FOXP4-AS1 predicts poor prognosis of MCL and accelerates its progression of MCL through the miR-423-5p/NACC1 pathway.[ Yang et al revealed that the lncRNA FOXP4-AS1 participates in the development and progression of osteosarcoma by downregulating LATS1 via binding to LSD1 and EZH2. Furthermore, overexpression of lncRNA FOXP4-AS1 led to enhanced proliferation, migration, and invasion; shortened the G0/G1 phase; and inhibited the cell cycle.[ The study of Zhong et al demonstrated that high expression of lncRNA FOXP4-AS1 in NPC portended poor outcomes. LncRNA FOXP4-AS1upregulated STMN1 by interacting with miR-423-5p as a competing endogenous RNA (ceRNA) to promote NPC progression.[ Wu et al confirmed that the lncRNA FOXP4-AS1 is activated by PAX5 and promotes the growth of prostate cancer by sequestering miR-3184-5p to upregulate FOXP4.[ Niu et al,s research found that lncRNA FOXP4-AS1, which is upregulated in esophageal squamous cell carcinoma (ESCC), promotes FOXP4 expression by enriching MLL2 and H3K4me3 in the FOXP4 promoter through a “molecular scaffold.” Moreover, FOXP4, a transcription factor of β-catenin, promotes the transcription of β-catenin and ultimately leads to malignant progression of ESCC.[ Liu et al found that lncRNA FOXP4-AS1 may function in pancreatic ductal adenocarcinoma (PDAC) by participating in biological processes and pathways, including oxidative phosphorylation, tricarboxylic acid cycle, and classical tumor-related pathways such as NF-kappaB and Janus kinase/signal transducers, in addition to activators of transcription, cell proliferation, and adhesion.[ Activation of DNA repair is one of the reasons for chemoresistance, and the myc-pathway has been associated with the acquisition of temozolomide resistance in glioblastoma through a c-Myc–miR-29c–REV3L network.[ Through pathway analysis, Huang et al suggested that the DNA repair/MYC gene set is enriched in low-grade glioma patients with high expression of lncRNA FOXP4-AS1. Therefore, overexpression of lncRNA FOXP4-AS1may may affect temozolomide the prognosis of cancer. However, this result needs to be further explored.[ It is noteworthy that the present study had some limitations. First, only English language reports have been considered, so we might have missed important studies published in other languages, and the studies included were all from China, and the results may best reflect the clinical characteristics of Asian cancer patients. Second, considering the relatively small number of samples in the literature, it is still necessary to investigate a single tumor type in a larger number of samples, and additional studies are needed to assess DFS and PFS. Due to this, potential publication bias is very likely to exist, despite the lack of evidence from our statistical tests. In the end, the HRs and 95%CIs were extracted by the indirect method, which was inevitably biased. Therefore, it is necessary to increase the number of high quality studies that contain a large number of samples to avoid the various factors in the compound.
We conducted the first systematic review and estimation of the relationship between abnormal lncRNA FOXP4-AS1 expression and survival and clinical outcomes in patients with tumors. Based on our results, high expression levels of lncRNA FOXP4-AS1 were associated with poor OS and DFS, making this gene a potential prognostic biomarker. Given the limitations of this study, a more large-scale, high-quality study on a variety of ethnic populations is necessary to assess the value of lncRNA FOXP4-AS1 in tumors.
Conceptualization: Qiang Shu, Fei Xie. Data curation: Xiaoling Liu, Jushu Yang. Formal analysis: Xiaoling Liu. Funding acquisition: Fei Xie. Investigation: Xiaoling Liu, Jushu Yang, Tinggang Mou. Methodology: Qiang Shu, Xiaoling Liu, Jushu Yang, Tinggang Mou. Project administration: Qiang Shu, Fei Xie. Software: Tinggang Mou. Supervision: Fei Xie. Writing – original draft: Qiang Shu. Writing – review & editing: Qiang Shu. | true | true | true |
|
PMC9592486 | Yuning Lin,Ying Li,Yongquan Chen,Zhongying Zhang | LncRNA ALMS1-IT1 is a novel prognostic biomarker and correlated with immune infiltrates in colon adenocarcinoma | 21-10-2022 | ALMS1-IT1,biomarker,LncRNA | Colon adenocarcinoma (COAD) is one of the most serious cancers. It is important to accurately predict prognosis and provide individualized treatment. Evidence suggests that clinicopathological features and immune status of the body are related to the occurrence and development of cancer. Expression of long non-coding RNA (LncRNA) ALMS1 intronic transcript 1 (ALMS1-IT1) is observed in some cancer types, and we believe that it may have the potential to serve as a marker of COAD. Therefore, we used the data obtained from the cancer genome atlas (TCGA) database to prove the relationship between ALMS1-IT1 and COAD. Wilcoxon rank sum test, Chi-square test, Fisher exact test and logistic regression were used to evaluate relationships between clinical-pathologic features and ALMS1-IT1 expression. Receiver operating characteristic curves were used to describe binary classifier value of ALMS1-IT1 using area under curve score. Kaplan–Meier method and Cox regression analysis were used to evaluate factors contributing to prognosis. Gene oncology (GO) and (Kyoto Encyclopedia of Genes and Genomes) KEGG enrichment analysis were used to predict the function of differentially expressed genes associated with ALMS1-IT1. Gene set enrichment analysis (GSEA) was used to predict canonical pathways associated with ALMS1-IT1.Immune infiltration analysis was performed to identify the significantly involved functions of ALMS1-IT1. Starbase database was used to predict miRNAs and RNA binding proteins (RBPs) that may interact with ALMS1-IT1. Increased ALMS1-IT1 expression in COAD was associated with N stage (P < .001), M stage (P = .003), Pathologic stage (P = .002), and Primary therapy outcome (P = .009). Receiver operating characteristic curve suggested the significant diagnostic and prognostic ability of ALMS1-IT1 (area under curve = 0.857). High ALMS1-IT1 expression predicted a poorer overall-survival (P = .005) and poorer progression-free interval (PFI) (P = .012), and ALMS1-IT1 expression was independently correlated with PFI in COAD patients (hazard ratio (HR) :1.468; 95% CI: 1.029–2.093; P =.034) (HR: 1.468; 95% CI: 1.029–2.093; P = .034). GO, KEGG, GSEA, and immune infiltration analysis showed that ALMS1-IT1 expression was correlated with regulating the function of DNA and some types of immune infiltrating cells. ALMS1-IT1 expression was significantly correlated with poor survival and immune infiltrations in COAD, and it may be a promising prognostic biomarker in COAD. | LncRNA ALMS1-IT1 is a novel prognostic biomarker and correlated with immune infiltrates in colon adenocarcinoma
Colon adenocarcinoma (COAD) is one of the most serious cancers. It is important to accurately predict prognosis and provide individualized treatment. Evidence suggests that clinicopathological features and immune status of the body are related to the occurrence and development of cancer. Expression of long non-coding RNA (LncRNA) ALMS1 intronic transcript 1 (ALMS1-IT1) is observed in some cancer types, and we believe that it may have the potential to serve as a marker of COAD. Therefore, we used the data obtained from the cancer genome atlas (TCGA) database to prove the relationship between ALMS1-IT1 and COAD. Wilcoxon rank sum test, Chi-square test, Fisher exact test and logistic regression were used to evaluate relationships between clinical-pathologic features and ALMS1-IT1 expression. Receiver operating characteristic curves were used to describe binary classifier value of ALMS1-IT1 using area under curve score. Kaplan–Meier method and Cox regression analysis were used to evaluate factors contributing to prognosis. Gene oncology (GO) and (Kyoto Encyclopedia of Genes and Genomes) KEGG enrichment analysis were used to predict the function of differentially expressed genes associated with ALMS1-IT1. Gene set enrichment analysis (GSEA) was used to predict canonical pathways associated with ALMS1-IT1.Immune infiltration analysis was performed to identify the significantly involved functions of ALMS1-IT1. Starbase database was used to predict miRNAs and RNA binding proteins (RBPs) that may interact with ALMS1-IT1. Increased ALMS1-IT1 expression in COAD was associated with N stage (P < .001), M stage (P = .003), Pathologic stage (P = .002), and Primary therapy outcome (P = .009). Receiver operating characteristic curve suggested the significant diagnostic and prognostic ability of ALMS1-IT1 (area under curve = 0.857). High ALMS1-IT1 expression predicted a poorer overall-survival (P = .005) and poorer progression-free interval (PFI) (P = .012), and ALMS1-IT1 expression was independently correlated with PFI in COAD patients (hazard ratio (HR) :1.468; 95% CI: 1.029–2.093; P =.034) (HR: 1.468; 95% CI: 1.029–2.093; P = .034). GO, KEGG, GSEA, and immune infiltration analysis showed that ALMS1-IT1 expression was correlated with regulating the function of DNA and some types of immune infiltrating cells. ALMS1-IT1 expression was significantly correlated with poor survival and immune infiltrations in COAD, and it may be a promising prognostic biomarker in COAD.
Colorectal cancer (CRC) includes colon adenocarcinoma (COAD) and rectal adenocarcinoma, according to pathological classification, approximately 80% to 90% of CRC cases are COAD.[ CRC ranks third in terms of incidence (10.2% of total cases) and is the second leading cause of cancer-related death globally (9.2% of all cases).[ In the past few decades, the incidence and mortality of CRC have steadily declined globally, however, CRC is still the most common gastrointestinal malignancy and the second largest cancer-related disease cause of death.[ The use of chemotherapy and surgical resection for malignant CRC is increasing, but the effect of these treatments has not been significantly improved. About half of CRC recur and patients die within 5 years.[ Therefore, it is necessary to identify new diagnostic, prognostic biomarkers and therapeutic targets, as well as to study the potential molecular mechanisms of COAD. Good diagnostic and prognostic biomarkers should be closely related to the prognosis of patients and easy to detect. Encouragingly, a large amount of evidence indicates that the regulatory role of long non-coding RNAs (LncRNAs) is related to the development and progression of a variety of cancers.[ LncRNAs are ≥200 nucleotides in length and do not encode proteins. According to its position and background in the genome, lncRNA can be divided into 5 main types: intergenic lncRNAs, intragenic lncRNAs, bidirectional lncRNAs, sense lncRNAs and antisense lncRNAs.[ The mechanisms of lncRNA regulating gene expression mainly include transcriptional repression, RNA-DNA interaction (chromatin remodeling), nuclear RNA-RNA interaction and cytoplasmic RNA-RNA interaction. Their functions are to regulate a series of cellular biological processes, including chromatin remodeling, transcriptional and post-transcriptional events.[ The most recognized molecular mechanism of lncRNAs is to act as a miRNA “sponge” to regulate downstream target genes.[ LncRNAs is abnormally expressed in various types of cancer cells and plays an important role in several common hallmarks of cancer.[ In addition, a growing number of studies indicate that lncRNAs may be identified as novel biomarkers for diagnosis, prognosis and metastasis prediction in various cancers.[ These year, experiments have demonstrated that several lncRNAs are CRC-specific lncRNAs, such as PCAT-1, RP11-462C24.1, HOTAIR, and MALAT1 as candidate diagnostic biomarkers.[ ALMS1-IT1,officical full name is ALMS1 intronic transcript 1. Up to now, there are few studies on ALMS1-IT1, and some studies believe that ALMS1-IT1 has prognostic value.[ Recent studies have predicted that the expression of ALMS1-IT1 may be related to ferroptosis.[ Bioinformatics analysis predicts that ALMS1-IT1 can serve as a prognostic biomarker for Head and neck squamous cell carcinoma (HNSCC).[ Experiments have shown that ALMS1-IT1/AVL9 promotes the malignant progression of lung adenocarcinoma (LUAD) in part by regulating the cyclin-dependent kinase pathway.[ Based on previous research results, we believe that ALMS1-IT1 may play an important role in the occurrence and development of COAD. Meanwhile, the role of ALMS1-IT1 in COAD has not been reported. Hence, in this research, we used the COAD RNA-seq data in the cancer genome atlas (TCGA) database to compare the difference of ALMS1-IT1 expression between tumor tissues and normal samples, and investigated the correlation between ALMS1-IT1 expression levels and clinical pathological features of COAD. Next, we evaluated the prognostic value of ALMS1-IT1 in COAD. In addition, gene oncology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) and gene set enrichment analysis (GSEA) were performed on the high and low expression groups of ALMS1-IT1 to reveal its possible functions. Meanwhile, Starbase database was used to predict miRNAs and RNA binding proteins (RBPs) that may interact with ALMS1-IT1. Finally, by analyzing the correlation between ALMS1-IT1 expression and immune infiltration, we comprehensively explored and discussed the potential mechanism of ALMS1-IT1 regulating the occurrence and development of COAD.
We used TCGA database (https://portal.gdc.cancer.gov/) to collect RNA-seq data and clinical information from 521 cases of COAD projects, including 41 cases with matched adjacent tissues. The downloaded data format was level 3 HTSeq-fragments per kilobase per million and then was converted into transcripts per million format for subsequent analysis. We also download transcripts per million format RNA-seq data in TCGA and Genotype-Tissue Expression database that uniformly processed by Toil process from University of California Santa Cruz Xena (https://xenabrowser.net/datapages/).[23] All procedures performed in this study were in accordance with the Declaration of Helsinki (as revised in 2013). We used R package (DESeq2)[ to go differential analysis of ALMS1-IT1 expression, adjusted P value < 0.05 and |logFC| > 2 were consider as cut off criteria, the Different Expression Genes (DEGs) obtained were used for GO, KEGG analysis, adjusted P value <.05 were consider as another cut off criteria, the DEGs obtained were used for GSEA. The R(version 3.6.0) package org. Hs.e.g..db(3.10.0) was used to conversion gene ID, cluster Profiler(3.14.3) was used to perform GO, KEGG, and GSEA between high- and low-ALMS1-IT1 groups.[ According to the default statistical method, the process was repeated 1000 times for each analysis and selected c2.cp.v7.2.symbols.gmt in MSigDB Collections as the reference gene collection, false discovery rate q-value < 0.25 and adjusted P adjust <.05 were considered to be significantly enriched.
The immune infiltration analysis of COAD was performed by single sample GSEA (ssGSEA) method from R (v.3.6.3) package GSVA (version 1.34.0),[ and we quantified the infiltration levels of 24 immune cell types from gene expression profile in the literature.[ In order to discover the correlation between ALMS1-IT1 and the infiltration levels of 24 immune cells, P values were determined by the Pearson and Wilcoxon rank sum test.
Starbase database (https://starbase.sysu.edu.cn/) was used to predict miRNAs and RBPs that may interact with ALMS1-IT1.
All statistical analyses were performed using R(v.3.6.3). Wilcoxon rank sum test, chi square test, Fisher exact test and logistic regression were used to analyze the relationship between clinical pathologic features and ALMS1-IT1. Kaplan–Meier method was used to calculate the overall survival rate and progression-free interval (PFI) of COAD patients from TCGA. Univariate and multivariate analysis were performed to estimate the association between clinical and genetic clinical characteristics and PFI using Cox proportional hazard models. P values <.05 were considered statistically significant.
This study does not involve experiments that require ethical approval.
In order to identify the difference of ALMS1-IT1 expression between COAD and normal tissues, we analyzed the expression level of ALMS1-IT1 in 480 COAD tissues and 41 adjacent normal colon tissues, and found that ALMS1-IT1 was highly expressed in COAD tissues (P < .001, Fig. 1A). Meanwhile, we also analyzed the expression of ALMS1-IT1 in 41 COAD tissues and their matched adjacent tissues. The results indicated that COAD tissues highly expressed ALMS1-IT1 (P < .001, Fig. 1B). Moreover, the expression of ALMS1-IT1 in normal samples of Genotype-Tissue Expression combined TCGA database and COAD samples of TCGA database was compared. We also found that ALMS1-IT1 was significantly overexpressed in COAD samples (P = .037, Fig. 1C). In addition, we used the receiver operating characteristic curve to analyze the effectiveness of ALMS1-IT1 expression level to distinguish COAD tissues from non-tumor tissues. The area under curve of ALMS1-IT1 was 0.857, suggesting that ALMS1-IT1 could be served as an ideal biomarker to distinguish COAD from non-tumor tissue (Fig. 1D). The characteristics of patients were shown in Table 1, in which 478 primary COAD with both clinical and gene expression data were collected from TCGA database. According to the mean value of relative ALMS1-IT1 expression, the patients with COAD were divided into high (n = 239) and low (n = 239) expression groups. The association between the expression level of ALMS1-IT1 and the clinicopathological characteristics of COAD patients was evaluated. Chi-square test or Fisher’s exact test revealed that ALMS1-IT1 expression was associated with N stage (P < .001), Gleason score (P = .002), primary therapy outcome (P = .001) and residual tumor (P < .001). Logistic regression method was also used to show the relationship between the clinicopathological characteristics of COAD and expression level of ALMS1-IT1. The results suggested that ALMS1-IT1 was significantly related to N stage (P < .001), M stage (P = .003), Pathologic stage (P = .002), and Primary therapy outcome (P = .009). Logistic regression method was also used to show the relationship between the clinicopathological characteristics of COAD and expression level of ALMS1-IT1. The results suggested that ALMS1-IT1 was significantly related to N stage (P < .001), M stage (P = .002), Pathologic stage (P < .001), primary therapy outcome (P < .001) (Table 2 and Fig. 2).
The association between ALMS1-IT1 expression and OS or PFI of patients with COAD was evaluated by Kaplan–Meier analysis, which indicated that expression of ALMS1-IT1 is positively correlated with poor OS (P = .005, Fig. 3A) and poor PFI (P = .012, Fig. 3B) of COAD patients.
GSEA was used to identify ALMS1-IT1-related signaling pathways. GSEA revealed significant differences (Padj < 0.05, false discovery rate < 0.25) in enrichment of MSigDB Collection (c2.cp.v7.2.symbols.gmt). We selected the top 9 data sets with high value of normalized enrichment score (Table 3 and Fig. 4).
To estimate the potential functions of DEGs in high-risk versus (vs) low-risk groups, we identify DEGs of ALMS1-IT1 in TCGA-COAD data under cutoff criteria of adjusted P value <.05 and |logFC|>2.KEGG pathway and GO annotation were performed by R package clusterProfiler(3.14.3).GO reveals the catalogs of biological process, cellular component, and molecular function.After multiple-test correction, KEGG pathways and GO terms with corrected P (P.adjust) value <.05 were considered to be prominently enriched in DEGs. We selected top 5 of the lowest adj. P value of GO and KEGG pathway enrichment analysis of 3303 DEGs related to ALMS1-IT1 in TCGA-COAD data (Table 4 and Fig. 5).
We further analyzed the correlation between expression of ALMS1-IT1 and immune infiltration by ssGSEA with Pearson. The results showed that the expression of ALMS1-IT1 was negatively correlated with most immune cells, and the top 3 negative correlation coefficients were natural killer (NK) cells, immature dendritic cell (iDC) and NK CD56bright cells (P < .001, Fig. 6).
Thirty-nine miRNAs and 42 RBPs that may interact with ALMS1-IT1 were identified using Starbase database (Fig. 7).
In this study, the expression of LncRNA ALMS1-IT1 in COAD and its correlation with COAD diagnosis and prognosis were explored. In general, lncRNAs exert regulatory functions at different levels of gene expression, including chromatin modification, transcription, and post-transcription.[ lncRNAs can interact with chromatin remodeling complexes to induce heterochromatin formation at specific genomic sites and reduce gene expression. In addition, lncRNAs interact with RNA-binding proteins and transcription factor co-activators, or regulate transcription by regulating the main promoters of their target genes. Mechanically, LncRNAs can communicate with DNA, mRNAs, ncRNAs and proteins and play cancer-related regulatory roles, such as signals, decoys, scaffolds and guidelines.[ In addition, lncRNAs were often involved in different stages of CRC, from precancerous polyps to distant metastasis, which can be regarded as potential effective diagnostic biomarkers.[ ALMS1-IT1 is a recently discovered lncRNA, which has been shown to play a key role in regulating tumor progression and predicting the survival time of tumor patients.[ Luan’s study points out that ALMS1-IT1 was highly expressed in LUAD, and the high expression of ALMS1-IT1 lead to poor prognosis in LUAD patients. Importantly, overexpression of ALMS1-IT1 helps to promote the viability of LUAD cells in vitro.[ Lei Y’s[ reveals the significance of the interaction between lncRNAs and ceRNAs in small cell lung cancer, indicating that the integration of expression profiles and alternative splicing can be used to identify biomarkers and potential pathological changes, and ALMS1-IT1 is one of the critical gene. Lu Xing et al[ reported that ALMS1-IT1 is up-regulated in high-risk groups of head and neck squamous cell carcinoma (HNSCC), which is related to the poor prognosis of HNSCC patients. In addition, it was also found that ALMS1-IT1 is a lncRNA targeting most miRNAs and proteins in HNSCC. All these studies suggest that ALMS1-IT1 may play different roles in various cancer types. The present study demonstrated the elevated level of ALMS1-IT1 in COAD tissues, which is associated with poor patient outcome. A highlight of this work is to predict the potential mechanisms by which ALMS1-IT1 regulates the development of COAD. Through GO and KEGG, ALMS1-IT1 related gene were found to be involved in nucleosome assembly, chromatin assembly and DNA complex formation, indicating that ALMS1-IT1 may play a role in cell replication. Through GSEA,ALMS1-IT1 was found related in DNA methylation and histone methylation, indicating that ALMS1-IT1 may play a role in the maintenance of cell metabolism and nucleic acid modification and protein modification. Another important aspect of this study is to investigate the relationship between ALMS1-IT1 expression and diverse immune infiltration levels in COAD. Our results revealed a moderate relationship between ALMS1-IT1 expression and infiltration level of NK cells, iDC and NK CD56 bright cells in COAD. These correlations could be indicative of a potential mechanism by which ALMS1-IT1 inhibits the function of NK cells, NK CD56 bright cells and iDC, subsequently promotes the function of T central memory, and thus exerts its inhibitory effect on COAD. To our knowledge, despite some limitations, this is the first work to explore the relationship between ALMS1-IT1 and COAD. First of all, the current research is mainly based on bioinformatics analysis, which can be further strengthened through experimental research. Second, the number of healthy subjects as controls is very different from the number of cancer patients. Finally, retrospective research still has its limitations, especially the inconsistent intervention measures and lack of relevant information. Therefore, follow-up studies are needed to further verify our findings.
Collectively, we observed increased ALMS1-IT1 in COAD, which was also related to poor OS and poor PFI. Moreover, ALMS1-IT1 might participate in the development of COAD via affecting the function of DNA and immune infiltrating cells. The current study partially unveiled the roles of ALMS1-IT1 in COAD and provided a potential biomarker for the diagnosis and prognosis of COAD. This study was supported by Xiamen medical and health guidance project (No.3502Z20209072). Xiantao Academic provides technical support for R analysis.
Formal analysis: Ying Li, Zhongying Zhang. Methodology: Yuning Lin, Ying Li, Yongquan Chen, Zhongying Zhang. Writing – original draft: Yuning Lin. Writing – review & editing: Yuning Lin, Ying Li, Yongquan Chen, Zhongying Zhang. | true | true | true |
|
PMC9592609 | Tian Zhou,Yuxin Li,Xiaoyu Li,Fanzhuo Zeng,Yanxia Rao,Yang He,Yafei Wang,Meizhen Liu,Dali Li,Zhen Xu,Xin Zhou,Siling Du,Fugui Niu,Jiyun Peng,Xifan Mei,Sheng-Jian Ji,Yousheng Shu,Wei Lu,Feifan Guo,Tianzhun Wu,Ti-Fei Yuan,Ying Mao,Bo Peng | Microglial debris is cleared by astrocytes via C4b-facilitated phagocytosis and degraded via RUBICON-dependent noncanonical autophagy in mice | 24-10-2022 | Cellular neuroscience,Microglia,Astrocyte | Microglia are important immune cells in the central nervous system (CNS) that undergo turnover throughout the lifespan. If microglial debris is not removed in a timely manner, accumulated debris may influence CNS function. Clearance of microglial debris is crucial for CNS homeostasis. However, underlying mechanisms remain obscure. We here investigate how dead microglia are removed. We find that although microglia can phagocytose microglial debris in vitro, the territory-dependent competition hinders the microglia-to-microglial debris engulfment in vivo. In contrast, microglial debris is mainly phagocytosed by astrocytes in the brain, facilitated by C4b opsonization. The engulfed microglial fragments are then degraded in astrocytes via RUBICON-dependent LC3-associated phagocytosis (LAP), a form of noncanonical autophagy. Interference with C4b-mediated engulfment and subsequent LAP disrupt the removal and degradation of microglial debris, respectively. Together, we elucidate the cellular and molecular mechanisms of microglial debris removal in mice, extending the knowledge on the maintenance of CNS homeostasis. | Microglial debris is cleared by astrocytes via C4b-facilitated phagocytosis and degraded via RUBICON-dependent noncanonical autophagy in mice
Microglia are important immune cells in the central nervous system (CNS) that undergo turnover throughout the lifespan. If microglial debris is not removed in a timely manner, accumulated debris may influence CNS function. Clearance of microglial debris is crucial for CNS homeostasis. However, underlying mechanisms remain obscure. We here investigate how dead microglia are removed. We find that although microglia can phagocytose microglial debris in vitro, the territory-dependent competition hinders the microglia-to-microglial debris engulfment in vivo. In contrast, microglial debris is mainly phagocytosed by astrocytes in the brain, facilitated by C4b opsonization. The engulfed microglial fragments are then degraded in astrocytes via RUBICON-dependent LC3-associated phagocytosis (LAP), a form of noncanonical autophagy. Interference with C4b-mediated engulfment and subsequent LAP disrupt the removal and degradation of microglial debris, respectively. Together, we elucidate the cellular and molecular mechanisms of microglial debris removal in mice, extending the knowledge on the maintenance of CNS homeostasis.
Microglia are resident immune cells in the central nervous system (CNS) that play vital roles in CNS development, homeostasis and disease. As professional phagocytes, microglia engulf cell corpses, excessive cells, invading pathogens and amyloid beta (Aβ). Efficient phagocytosis is essential for maintenance of the homeostasis of the CNS. Different from neurons that are lifelong surviving in the adulthood, yolk sac-derived microglia undergo constant turnover. During this process, old microglia die, and new microglia regenerate via self-renewal. The generation of newly formed microglia is coupled with the apoptosis of dysfunctional microglia to maintain the stability of the cell population. In the homeostatic CNS, approximately 30% of microglia turn over each year in both mouse and human brains. In other words, approximately one-third of microglia die in the homeostatic brain each year. The speed of microglial turnover in diseased CNS is even faster than that in healthy CNS. If the cellular corpses are not removed in a timely manner, the accumulated debris can interfere with CNS function. Although microglia are the major cells responsible for the removal of cell debris, microglia per se are unlikely to phagocytose their congeneric corpses. In an extreme condition of rapid microglial depletion, more than 95% of CNS microglia are ablated within a few days. However, we and other research groups have found that rapid microglial depletion does not result in massive microglial debris accumulation or detectable inflammatory responses. Furthermore, even in the retina, where all microglia are pharmacologically depleted, no massive accumulation of microglial debris has been observed. The residual microglia are thus unable to rapidly clear debris due to their low number or even absence. This convergent evidence suggests an unknown yet efficient machinery that governs microglial debris removal. Therefore, investigating the removal of microglial debris is important for understanding the maintenance of CNS homeostasis. Nevertheless, the clearance of microglial debris is almost neglected, probably due to the primary scavenger role of microglia per se. To this end, we studied the cellular and molecular mechanisms underlying microglial debris removal. We first screened the major cell types in the brain as potential candidates, including astrocytes, pericytes, endothelial cells, vascular smooth muscle cells (VSMCs), oligodendrocyte precursor cells (OPCs), oligodendrocytes, neurons, neural stem cells (NSCs), microglia and CNS border-associated macrophages (BAMs). We found that pericytes, endothelial cells, VSMCs, OPCs, oligodendrocytes, neurons, NSCs and BAMs do not engulf microglial debris in vivo. Instead, microglial debris is primarily removed by astrocytes, a nonprofessional phagocyte, with a noncanonical and sophisticated reactive state. The astrocyte phagocytosis rates are positively correlated with the speed of microglial turnover under both physiological and pathological conditions. Although in vitro experiments indicate that microglia are able to engulf their congeneric debris, a territory-dependent competition of astrocytes hinders microglia-to-microglial debris engulfment in vivo. We also demonstrated that the astrocytic engulfment of microglial debris is facilitated through a C4b opsonization mechanism, and the engulfed microglial debris is subsequently degraded via LC3-associated phagocytosis (LAP), a form of noncanonical autophagy. The debris-containing LC3-associated phagosomes (LAPosomes) are then fused with lysosomes to form phagolysosomes for further degradation. The interference of LAP formation disrupts the degradation of engulfed microglial debris and induces accumulation of nondegraded debris in astrocytes, which indicates the essential role in the maintenance of CNS homeostasis. Collectively, our results uncover the phagocytic role of astrocytes in the scavenging of microglial debris. As nonprofessional phagocytes, astrocytes reportedly engulf neuronal debris. In this study, we extended this knowledge to the neglected facet of microglial debris clearance. Our study further identified the underlying mechanisms of astrocytic phagocytosis and intracellular degradation of microglial debris. Efficient removal by astrocytes avoids the accumulation of microglial debris. In conclusion, nonprofessional phagocytes phagocytose the debris of professional phagocytes via noncanonical autophagy for degradation. During this process, astrocytes exhibit a noncanonical reactivation phenotype. This study sheds new light on the maintenance of CNS homeostasis.
We first questioned which CNS cells are able to engulf microglial debris. In the homeostatic brain, microglial turnover is relatively slow, approximately 0.1% per day. The low turnover rate to some extent hinders researchers from studying the clearance of microglial debris. To overcome this obstacle, we utilized PLX5622, a colony-stimulating factor 1 receptor (CSF1R) inhibitor, to accelerate microglial cell death. We administered a PLX5622-formulated AIN-76A diet (1.2 g PLX5622 per kilogram of diet; referred to hereafter as PLX5622) or an AIN-76A control diet (CD) for 2 days to CX3CR1+/GFP mice, in which almost all microglia expressed GFP (Supplementary Fig. 1a). The dying microglia became collapsed and fragmented upon PLX5622 administration (Supplementary Fig. 2). Using this accelerated cell death model to facilitate the investigation of microglial debris removal, we detected evident GFP+ microglial debris incorporated with the astrocyte marker GFAP in the PLX5622-treated brain (Supplementary Fig. 1b). GFAP is an intermediate filament protein that does not illustrate the whole morphology of astrocytes. To better visualize the morphology and rigidly observe the colocalization, we crossed ALDH1L1-CreER mice, a brain and spinal cord astrocyte-specific mouse line, with Ai14 mice to obtain the ALDH1L1-CreER::Ai14 line. Tamoxifen was applied to induce tdTomato expression in brain and spinal cord astrocytes. Two weeks after the final dose, PLX5622 was administered for two days to kill CNS microglia (Fig. 1a). Consistent with our GFAP observations, a high amount of IBA1+ microglial debris was colocalized with tdTomato+ astrocytes in the brain (Fig. 1b–d and Supplementary Fig. 3). In addition, we intravenously introduced the blood‒brain barrier (BBB)-permeable AAV-PHP.eB Gfap-mCherry, which targets brain astrocytes and drives the fluorescent reporter to fully trace astrocytic morphology, to C57BL/6 J mice (Supplementary Fig. 4a). Thirty days after transduction, the astrocyte-labeled mice were administered PLX5622 for 2 or 4 days to kill brain microglia (Supplementary Fig. 4a). Similar to the observations from ALDL1L1-CreER::Ai14 mice, GFP+ microglial debris was colocalized with mCherry+ astrocytes in the brain (Supplementary Fig. 4b-c and Supplementary Fig. 5). The CNS includes the brain, spinal cord and retina. We thus asked whether astrocytes in the spinal cord and retina are also able to engulf microglial debris. We found that spinal cord astrocytes exhibited a similar capability of engulfing microglial debris (Fig. 1b–d and Supplementary Fig. 3). The retina contains two counterparts of brain astrocytes: Müller glia and retinal astrocytes. Müller glia vertically cross the whole retinal layers with their cell bodies localized in the inner nuclear layer (INL), whereas retinal astrocytes horizontally cross the retinal ganglion cell layer (GCL). ALDH1L1-CreER::Ai14 mice specifically labeled Müller glia but not retinal astrocytes (Supplementary Fig. 6a–c). We found that both Müller glia (visualized by tdTomato) and retinal astrocytes (visualized by GFAP) were capable of removing the debris of retinal microglia upon CSF1R inhibition (Fig. 1b–d and Supplementary Fig. 3). Therefore, the microglial debris in the spinal cord is removed by astrocytes, whereas that in the retina is removed by Müller glia and astrocytes. To further confirm the astrocytic engulfment of microglial debris in vitro, we prepared microglial debris by repeated cycles of freezing and thawing in liquid nitrogen to obtain dead microglia. The cell corpses were then labeled with pHrodo, a fluorescent staining dye, at 37 °C for 10 min After that, the pHrodo-labeled debris was incubated with purified astrocytes in complete medium (Supplementary Fig. 7) for 24–72 h (Fig. 1e). The pHrodo-labeled microglial debris was internalized in cultured astrocytes (Fig. 1f–g), and accumulation of microglial debris was detected from 24 to 72 h (Fig. 1g). The net amount of astrocyte-containing debris equals the phagocytosed debris minus the degraded debris. Because excessive amounts of microglial debris were added to the in vitro cell culture system, debris accumulation indicates that the speed of phagocytosis is faster than the speed of intracellular degradation. Thus, microglial debris accumulated in astrocytes. Collectively, this evidence from both in vivo and in vitro experiments demonstrates that astrocytes are able to engulf microglial debris. In contrast, GFP+ microglial debris was barely detected in pericytes (PDGFR-β), endothelial cells (CD31), VSMCs (α-SMA), oligodendrocytes (MBP and CC1) or neurons (TUJ1) (Fig. 2a, b). Although some OPCs (PDGFR-α) were able to phagocytose microglial debris, the engulfed debris content was limited (Fig. 2a, b). Furthermore, by using NESTIN-GFP transgenic mice (aka Nes-GFP mice), in which NSCs express GFP, we found that NSCs did not engulf microglial debris (Fig. 2b). In addition, we did not detect a compromised blood‒brain barrier (BBB) on days 2 and 4 after microglia depletion via 10-kDa dextran (Fig. 2c, d), which indicated that microglial debris is unlikely to be engulfed by infiltrating blood cells (e.g., monocytes). Together, our data indicate that upon microglial ablation, microglial debris is primarily scavenged by astrocytes and retinal Müller glia rather than pericytes, endothelial cells, VSMCs, OPCs, oligodendrocytes, NSCs, neurons or infiltrating blood cells.
Whether astrocytic engulfment is relevant to biological functions or occurs in forced microglial depletion is unclear. If astrocytic engulfment plays vital roles in the maintenance of CNS homeostasis, it should be positively correlated with the speed of microglial turnover. We thus explored the biological significance of astrocytic engulfment under two scenarios. First, microglia are heterogeneous cells with variable turnover rates in different brain regions under physiological conditions. We thus examined the astrocytic engulfment of microglial debris under physiological conditions in brain regions with different microglial turnover speeds, including the cortex (relatively slow), hippocampus (relatively fast) and olfactory bulb (relatively fast). Consistent with the regional differences in microglial turnover speeds, the astrocytes in the cortex possessed a relatively low level of GFP+ debris, whereas the levels of internalized microglial debris in the hippocampus and olfactory blub astrocytes were relatively high (Fig. 3a–c and Supplementary Fig. 8). Second, microglial turnover is accelerated in plaque-associated brain regions in Alzheimer’s disease (AD). We then tested the astrocytic engulfment of microglial debris in 5xFAD mice, a mouse model of AD, at 8 and 20 months of age (Fig. 3d) and found that astrocytes in plaque proximal regions exhibited higher capacities for microglial debris engulfment than those in plaque distal regions (Fig. 3e–f). Therefore, astrocytic engulfment orchestrates microglial turnover under physiological and pathological conditions, which indicates its biological relevance and suggests a critical role in the maintenance of CNS homeostasis.
Microglia are professional phagocytes in the CNS. We thus further asked whether microglia are able to remove their congeneric cell debris. The major obstacle to studying microglia-to-microglia engulfment is distinguishing engulfed microglial debris from engulfed microglia in vivo. To this end, we designed a sparse labeling system by treating CX3CR1-CreER::Ai14 mice with a low dose of tamoxifen (Fig. 4a). Thirty days after tamoxifen treatment, approximately 54% of microglia were labeled with tdTomato (Fig. 4a–c), and the animals were then administered PLX5622 to partially deplete microglia (Fig. 4a). Two days after depletion, approximately 51% of microglia were tdTomato+, and this value is comparable to the percentage before PLX5622 treatment (Fig. 4c). The results indicate that tdTomato+ and tdTomato− microglia are identically ablated upon PLX5622 treatment. If surviving microglia are capable of engulfing microglial debris (e.g., tdTomato− microglia engulf tdTomato+ debris), tdTomato+ debris should be observed in tdTomato− microglia. However, tdTomato+ debris was not detected in tdTomato− microglia (Fig. 4b, c), which suggests that the surviving microglia do not phagocytose microglial debris. On the one hand, CX3CR1-CreER::Ai14 (Cx3cr1wt/ko-CreER;Ai14wt/mut) mice are Cx3cr1-haploinsufficient. On some occasions, Cx3cr1-haploinsufficient microglia might exhibit a nonidentical phenotype to wild-type microglia. To avoid its potential influence and more rigidly confirm this conclusion, we generated P2Y12-CreER-GFP mice through in situ insertion of P2A-CreERT2-P2A-GFP between the last exon and the 3ʹ-UTR of the P2ry12 gene (Supplementary Fig. 9a). Almost all microglia expressed GFP, whereas non-myeloid cells in the brain were negative for GFP (Supplementary Fig. 9b-c). We thus sparsely labeled microglia in P2Y12-CreER-GFP::Ai14 mice and ablated microglia with PLX5622 (Fig. 4a). Similarly, tdTomato+ debris was still not detected in tdTomato− microglia of P2Y12-CreER-GFP::Ai14 mice (Fig. 4d, e), which further confirmed that the surviving microglia do not phagocytose microglial debris. On the other hand, the surviving microglia might be dysfunctional upon PLX5622 administration, which may not reflect the debris engulfment capability under physiological conditions. To address this question, we also compared the engulfment of debris under physiological conditions between sparsely labeled CX3CR1-CreER::Ai14 and P2Y12-CreER-GFP::Ai14 mice with ALDH1L1-CreER::Ai14 mice before CSF1R inhibition (Figs. 4a and 3a). Even under physiological conditions (with a relatively low microglial turnover rate), microglia debris-containing astrocytes were observed (Fig. 4f, 4.552%). In contrast, no obvious tdTomato+ debris was detected in tdTomato− nondysfunctional microglia (Fig. 4f, 0.036% in CX3CR1-CreER::Ai14 mice and 0.229% in P2Y12-CreER-GFP::Ai14 mice, adjusted by the sparse labeling efficiency). Therefore, our results indicate that microglial debris is not primarily removed by microglia. Notably, compared with the microglial debris engulfment rate of microglia (almost zero), the astrocytic engulfment rate under physiological conditions (Fig. 4f) further confirmed its physiological relevance in microglial debris removal. Next, we asked why microglia, the primary scavenger of the CNS, do not phagocytose their congeneric debris in the brain. We thus tested the capability of microglia-to-microglial debris phagocytosis in vitro. Primary microglia were harvested and cocultured with pHrodo-labeled microglial debris for 24 h (Fig. 4g). Unexpectedly, the pHrodo-labeled microglial debris was phagocytosed by the cultured microglia (Fig. 4h, i), which suggested that microglia retain the capability of engulfing their congeneric debris. Provided that our in vivo sparse labeling did not observe obvious microglia-to-microglial debris phagocytosis in physiological condition or upon microglial depletion (Fig. 4a–f), we thus concluded a territory-dependent competition model (hypothesis): microglia and astrocytes competitively phagocytose microglial debris. Under normal conditions, microglia evenly tile the brain and do not invade the territory of their congeneric neighbors (Fig. 4j (I)). In contrast, the territories between astrocytes and microglia overlapped (Fig. 4j (I)). When a microglial cell dies, the cell locally collapses (Fig. 4j (II)). The territory-overlapping astrocytes are close to dead microglia (Fig. 4j (II); cyan hexagons), whereas neighboring microglia are relatively distal (Fig. 4j (II); red microglia). Microglial debris is first removed by overlapping astrocytes before spreading to distal microglia for phagocytosis. Therefore, even though microglia can engulf their congeneric debris, microglial debris is competitively removed by spatial proximal astrocytes before coming into contact with relatively distal microglia. As a result, microglial debris is primarily removed in vivo by astrocytes rather than microglia. This territory-dependent competition model echoes previous observations of olfactory nerve debris clearance by olfactory ensheathing cells, in which little phagocytosis by distal microglia/macrophages was detected.
Brain myeloid cells comprise microglia and BAMs. Our P2Y12-CreER-GFP mouse can distinguish parenchymal microglia (GFP+) from meningeal, choroid plexus and perivascular macrophages (GFP-) (Supplementary Fig. 9b, c). To test whether BAMs are able to phagocytose microglial debris, we utilized tamoxifen to induce microglial cell death without affecting BAMs in P2Y12-CreER-GFP::DTA mice (Fig. 5a). Five days after tamoxifen administration, 30.10% of microglia were ablated. However, meningeal, choroid plexus or perivascular macrophages (F4/80+ GFP−) did not engulf GFP+ debris upon microglial depletion (Fig. 5b). Therefore, microglial debris is not removed by BAMs.
By immunostaining (in vivo), enzyme-linked immunosorbent analysis (ELISA; magnetic-activated cell sorting (MACS)-sorted astrocytes) and quantitative PCR (qPCR; MACS-sorted astrocytes), we observed robust upregulation of GFAP in astrocytes during microglial depletion (Fig. 6a–c). We and other research groups previously demonstrated that microglial depletion does not induce inflammatory responses; thus, astrocytes are not directly influenced by inflammatory microenvironments upon PLX5622 treatment. In addition, astrocytes do not express CSF1R and thus cannot be directly influenced by PLX5622. Therefore, the alteration of astrocytes from PLX5622-treated mice represents the response of astrocytes to the phagocytosis of microglial debris. The upregulation of GFAP in PLX5622-treated mice suggests that astrocytes exhibit a reactive-like phenotype during microglial debris clearance. To understand how astrocytes respond to microglial debris engulfment, we utilized fluorescence-activated cell sorting (FACS) to harvest tdTomato+ astrocytes from tamoxifen-treated ALDH1L1-CreER::Ai14 mice and compared the gene profiles upon CD and PLX5622 administration by RNA sequencing (RNA-seq) (Fig. 6d; gating strategy in Supplementary Fig. 10a). The RNA-seq data were validated by qPCR (Supplementary Fig. 10b-c). Principal component analysis (PCA) revealed that the transcriptome characteristics of normal astrocytes (CD group) were different from those of astrocytes upon PLX5622 administration (Fig. 6e). During microglial debris engulfment, 542 genes were upregulated, whereas 459 genes were downregulated (definition of differentially expressed genes: |fold change | >= 2 and FDR < = 0.01) (Fig. 6f). We then investigated the reactive astrocyte markers of the reactive astrocyte consensus statement. Most of the reactive astrocyte markers tended to exhibit differential expression (Fig. 6g, h). Among the positively correlated reactive astrocyte markers, only 8 out of 20 genes were upregulated (Fig. 6g, h), and 12 out of 20 genes were downregulated or not significantly changed (Fig. 6g, h). In contrast, among the negatively correlated reactive astrocyte markers, an upregulation trend was observed (Fig. 6g, h). Therefore, the RNA-seq results suggest that astrocytes exhibit a noncanonical and sophisticated reactive phenotype when engulfing microglial debris.
We asked how astrocytes engulf microglial debris. Complement components play key roles in the phagocytosis process; thus, we first examined whether complement opsonization mediates astrocytic engulfment of microglial debris. Compared with the results obtained using culture medium with complement component-containing serum (FBS), the engulfment of microglial debris by astrocytes cultured in serum-free medium (without FBS) was significantly suppressed (Fig. 7a, b). We subsequently tested the potential role of C1q, C2, C3, C4a and C4b, the major complement components in the CNS, by both supplementation and preopsonization. We found that C4a and C4b were able to restore the engulfment capability of astrocytes, whereas C1q, C2 or C3 were not (Fig. 7a, b). Astrocytes increase their cellular processes in serum-free medium over time, as we showed that astrocytes cultured for an extended period of 14 days in serum-free medium display a significant number of processes (Supplementary Fig. 11a), mimicking the astrocyte morphology in vivo. We further confirmed that in serum-free astrocytes cultured for 14 days, C4b still facilitates the astrocytic engulfment of microglial debris (Supplementary Fig. 11b). Similar results were detected in the serum-free culture medium (Supplementary Fig. 12a, b). In addition, we found that microglial debris colocalized with C4 both in vivo and in vitro (Supplementary Fig. 12c–e), which indicated that microglial debris is opsonized by C4 upon microglial death and fragmentation. Next, we reanalyzed the expression levels of C1qa, C2, C3, C4a and C4b in microglia-depleted brains using our previously published RNA sequencing (RNA-seq) database of bulk tissue from whole brain homogenates. We found that C2, C3 and C4a were expressed at relatively low levels in the brain and that the levels did not changed upon microglial ablation (Fig. 7c, d). Although C4a is able to facilitate debris engulfment in vitro, it is unlikely to play a major role in the brain due to its low expression level. In contrast, we observed a significant upregulation of C4b (Fig. 7c, d), the opsonin that facilitates cell debris clearance by professional phagocytes. We double confirmed the upregulation of C4b in MACS-sorted astrocytes, one of the major sources, during microglial depletion by qPCR (Fig. 7e). Converging evidence thus suggests that the astrocytic engulfment of microglial debris is facilitated by C4b in vivo. To test this hypothesis in vivo, we generated C4b−/− mice by CRISPR/Cas9-mediated genome editing (Supplementary Fig. 13) and utilized F0 founder mice (either biallelic or multiallelic mosaic), which were verified lacking C4 expression (Supplementary Fig. 14). Compared with the results found for the wild-type, the level of engulfed microglial debris was significantly reduced in C4b−/− astrocytes under physiological conditions (except for the cortex, most likely due to the low turnover rate in this brain region) (Fig. 7f–h; D21) and upon microglial depletion (Fig. 7f–h; D23). Therefore, our results indicate that complement C4b opsonizes microglial debris and facilitates astrocytic engulfment in vivo. We detected nearly complete ablation of C1qa (Fig. 7c, d), C1qb and C1qc (refer to GSE108269) during microglial depletion. Microglia are the major source of C1q; therefore, this downregulation should be attributed to massive microglial depletion. Because C1q facilitates the phagocytosis process as an “eat-me” signal, its downregulation during microglial depletion suggests that C1q does not participate in microglial debris clearance. To fully confirm this conclusion in vivo, we examined the engulfment of microglial debris in C1qa−/− mice (Supplementary Fig. 15a). We found that C1q deficiency did not affect the astrocytic engulfment of microglial debris under physiological conditions or upon microglial depletion (Supplementary Fig. 15b-c), confirming the C1q-independent mechanism. This finding also indicates that C4-mediated microglial removal is not a compensation of C1q absence.
We subsequently explored the mechanism through which the engulfed microglial debris is degraded in astrocytes. We found that genes that were differentially expressed in microglia-depleted brains were enriched in autophagy among the biological process terms, as determined by the Gene Ontology (GO) analysis (gene ratio: 13/371, Q value: 0.0287). The autophagy machinery was recently found to be coupled with phagocytosis, in which LC3 associates with phagosomes to facilitate phagosome fusion with lysosomes and enhance the degradative efficiency, and this noncanonical autophagy process is termed LAP. We found that LC3 is associated with intra-astrocytic microglial particles (Fig. 8a, b), which suggests that the internalized debris is degraded through autophagy. The diameter of autophagosomes in canonical autophagy is typically 0.5–1 μm. In contrast, the LAPosomes in LAP, a noncanonical autophagy, typically have diameters larger than 1 μm. Based on our observations, the diameter of intra-astrocytic microglial fragment-containing LC3+ organelles was 1.52 ± 0.98 μm, significantly larger than the diameter of autophagosomes (0.42 ± 0.18 μm) in starvation-induced canonical autophagy (Fig. 8b–d). This finding indicates that microglial debris is degraded through the LAP mechanism. We further characterized the membrane ultrastructure by transmission electron microscopy (TEM). Different from the double-membrane autophagosome of canonical autophagy, the LAPosomes of noncanonical autophagy form a single membrane. Our results showed that the debris-containing single-membrane structures (LAPosomes of noncanonical autophagy) were markedly increased in microglial debris-engulfing astrocytes (Fig. 8e–f), which further confirmed the noncanonical autophagy mechanism. Another discrepancy between canonical and noncanonical autophagy is the molecular mechanism governing these two processes. Despite sharing some core machineries with canonical autophagy, LAP is uniquely dependent on RUBICON, a negative regulator of canonical autophagy. When RUBICON is knocked out or knocked down (Fig. 8g), astrocytes failed to form LAPosomes encapsulating the microglial debris but showed increased autophagosome puncta formation (Fig. 8h). As a consequence, the nondegraded microglial debris accumulated in RUBICON-knockout (Fig. 8k–m, Supplementary Fig. 16a–d) or RUBICON-knockdown astrocytes (Supplementary Fig. 16e–h). In addition, the levels of neutral lipids, the degradative product of cell debris, were significantly reduced in RUBICON-knockout or RUBICON-knockdown astrocytes (Fig. 8i–j). Therefore, the intra-astrocytic debris is degraded via RUBICON-dependent LAP. The intra-astrocytic microglial fragments colocalized with the lysosomal marker LAMP1 in PLX5622-treated mice (Fig. 9a, b) and primary astrocyte cultures (Fig. 9c, d). In addition, both GO and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses revealed that “lysosome” was enriched in astrocytes of the microglial depletion brain at day 2 (GO: gene ratio: 49/969, Q value: 1.47 × 10−8; KEGG: 20/481, Q value: 1.81 × 10−3). These results suggest that the debris-containing LAPosomes fused with lysosomes to form phagolysosomes for debris degradation. Indeed, when chloroquine was used to interfere with the formation of phagolysosomes, the level of neutral lipids, which are the degradation product, was significantly reduced in astrocytes (Fig. 9e). Thus, microglial debris-containing LAPosomes fuse with acidic lysosomes to form phagolysosomes for further degradation. Collectively, our results demonstrate that phagocytosed microglial debris is degraded through the LAPosome-to-phagolysosome axis in astrocytes. In summary, microglial debris is primarily engulfed by astrocytes, and this process is facilitated by C4 opsonization. The phagocytosed debris is then degraded in astrocytes via RUBICON-dependent LAP, and microglial debris-containing LAPosomes then fuse with lysosomes to form phagolysosomes, in which the debris is degraded into lipid droplets (Fig. 10).
As professional phagocytes and scavengers in the CNS, microglia are the major players engulfing cell corpses, excessive dendritic spines and Aβ aggregates. However, the identity of the cells that scavenge corpses of the professional scavenger is largely neglected. Importantly, even when microglia are massively and rapidly killed, the cell debris can be removed in a timely manner to avoid cellular debris accumulation. These findings imply the biological significance of microglial debris clearance in the maintenance of CNS homeostasis and function, but the cellular and molecular mechanisms remain elusive. Microglia do not participate in the removal of congeneric microglial debris in vivo (at least not the major scavenger). Furthermore, we excluded the participation of pericytes, endothelial cells, VSMCs, OPCs, oligodendrocytes, NSCs, neurons and infiltrating blood cells. Although the phagocytic capacities of microvascular endothelial cells, OPCs and neural stem (progenitor) cells have been reported by us and other groups, these cells are not extensively involved in microglial debris clearance. Instead, we found that microglial debris is primarily removed by astrocytes. Although professional phagocyte microglia can engulf nonprofessional phagocyte astrocytes, the removal of debris from these professional phagocytes is conducted by the nonprofessional phagocytes. Furthermore, we investigated the cellular and molecular mechanisms. Astrocyte engulfment is mediated by the opsonization of C4. The phagocytic cargos are then degraded via LAP, a form of noncanonical autophagy. As glial cells and nonprofessional phagocytes in the CNS, astrocytes reportedly clear the debris of dysfunctional neurons in brain injury and neurodegenerative disorders. A recent study demonstrated that astrocytes and microglia coordinate to clear cell corpses: microglia phagocytose cell bodies and apical dendrites of dying neurons, whereas astrocytes engulf numerous small dendritic apoptotic bodies. When Irf8 deficiency disables microglial phagocytosis, astrocytes can play a complementary role. Although these studies focused on delineating astrocytic phagocytosis, little attention has been given to the astrocytic contribution to microglial turnover, which has been partially attributed to the professional phagocytic role of microglia per se. Therefore, our study reveals the almost neglected yet important role of astrocytes in microglial debris clearance. In addition, although previous studies reported the phagocytic roles of astrocytes, the mechanism through which internalized cargos are routed to the degradation machinery remains unclear. Our current findings show how internalized cargos are degraded in astrocytes and thus offer insights into this nonprofessional phagocytic cell. In the homeostatic brain, the microglial turnover (death) rate is approximately 0.05 to 0.2% per day. In contrast, approximately 15% of cells are ablated within the first 24 h upon PLX5622 administration, and this value is 75 to 150 times faster than that under normal conditions. Even if numerous microglia are ablated within a relatively short period, the cell fragments are not evidently accumulated. In addition, microglia can respond to neural injury and migrate toward the injury site within 1 h. Even with such a rapid migration speed, microglia do not largely engulf their neighboring microglial debris upon territory-dependent competition of astrocytes in vivo. Thus, rapid cellular corpse removal indicates an efficient, even excessive, phagocytic capacity of astrocytes. The excessive astrocytic capability implies the significance of microglial debris removal that guarantees the microenvironment not affected. In contrast, microglia show accelerated cell death and/or turnover in CNS disorders. For instance, in ischemic stroke, microglia show substantial proliferation at the early stage. The cell number is then steeply reduced due to microglial cell death. In addition, plaqueproximal microglia exhibit a several-fold higher turnover rate than plaque distal microglia in AD. Thus, the excessive phagocytic ability of astrocytes ensures the timely removal of microglial debris, which confines a secondary injury and dampens CNS dysfunction. Collectively, the excessive capacity of phagocytosis by astrocytes might be critical to protect the CNS from the influence of accumulated cell debris thereby maintaining the CNS homeostasis.
All mouse experiments were conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee of Fudan University, Shenzhen Institute of Advanced Technology at Chinese Academy of Sciences and Shanghai Mental Health Center at Shanghai Jiao Tong University School of Medicine. This study was approved by the Institutional Animal Care and Use Committee of Shenzhen Institute of Advanced Technology at Chinese Academy of Sciences (SIAT-IACUC-190312-YGS-PB-A0576-01) and the Institutional Animal Care and Use Committee of Fudan University (2021JS ITBR-002).
CX3CR1GFP/GFP (B6.129P-Cx3cr1tm1Litt/J, Stock No: 005582), CX3CR1-CreER (B6.129P2(C)-Cx3cr1tm2.1(cre/ERT2)Jung/J, Stock No: 020940), Ai14 (B6. Cg-Gt(ROSA)26Sortm14(CAG-tdTomato)Hze/J, Stock No: 007914), 5xFAD (B6. Cg-Tg(APPSwFlLon,PSEN1*M146L*L286V)6799Vas/Mmjax, Stock No: 34848-JAX), DTA (B6.129P2-Gt(ROSA)26Sortm1(DTA)Lky/J, aka Rosa26-DTA), C1qa−/− (B6(Cg)-C1qatm1d(EUCOMM)Wtsi/TennJ) and RUBICON−/− (C57BL/6-Rubcnem1Dgre/J) mice were purchased from The Jackson Laboratory. RUBICONfl/fl (B6/JGpt-Rubcnem1Cflox/Gpt) mice were purchased from GemPharmatech. NESTIN-GFP (Nes-GFP) mice were procured from Cyagen Biosciences China. C57BL/6 J mice were purchased from Charles River (Beijing Vital River Laboratory Animal Technology). ALDH1L1-CreER mice were kindly donated by Prof. Tian-Ming Gao at Southern Medical University. C4b−/− mice were generated by CRISPR/Cas9-mediated genome editing. sgRNAs were designed using CHOPCHOP. Two selected sgRNAs targeting exons 1 and 2 of all four C4b transcripts, namely, GCTCCTCTGGGGGCTGGCCT and GCTGTACCCCCACCGACAGG were constructed by fusion of a T7 promoter and tracrRNA via overlapping PCR. Full-length sgRNA was produced and transcribed. The transcribed sgRNA and Cas9 mRNA were both purified. The mixture of sgRNA (25 ng/μL each) and Cas9 mRNA (50 ng/μL) was microinjected into the cytoplasm of zygotes of C57BL/6 mice to generate C4b−/− mice. Fifty-five founder mice (F0) were born and analyzed for CRISPR-edited indels. Mouse tail clips were used for PCR amplification with primers flanking the sgRNA regions ACGCACATGCACAGGGACAC and TCAAGGCTGAGCAGCACAAA. The amplicons were directly Sanger sequenced (Supplementary Data 1) and then subjected to CRISPR edit analysis of C4b indels using the Synthego ICE CRISPR Analysis Tool (https://www.synthego.com/products/bioinformatics/crispr-analysis). Among the 55 F0 founder lines generated in this study, 8 were identified as C4b indel mutants with >83% frameshift indels, including 5 biallelic lines (either homozygotes or compound heterozygotes) and 3 multiallelic mosaics. P2Y12-CreER-GFP (P2ry12-p2A-CreER-p2A-EGFP) mice were generated by Beijing Biocytogen through CRISPR/Cas9-based extreme genome editing (EGE) according to the authors’ design as a “fee-for-service”. The p2A-CreERT2-p2A-EGFP cassette was inserted between the last exon of the P2ry12 gene (transcript 201), the gene encoding P2Y12, and its 3ʹ UTR. ALDH1L1-CreER mice were crossed with Ai14 mice to obtain ALDH1L1-CreER::Ai14 (ALDH1L1+/CreER::Ai14wt/mut) mice. ALDH1L1-CreER mice were crossed with Ai14 and RUBICONfl/fl mice to obtain ALDH1L1-CreER::Ai14::RUBICONfl/fl (ALDH1L1-CreER+/CreER::Ai14 wt/mut::RUBICONfl/fl) mice, aka RUBICON-cKO mice. CX3CR1-CreER mice were crossed with Ai14 mice to obtain CX3CR1-CreER::Ai14 (CX3CR1+/CreER::Ai14wt/mut) mice. P2Y12-CreER-GFP mice were crossed with Ai14 mice to obtain P2Y12-CreER-GFP::Ai14 (P2Y12+/CreER-GFP::Ai14wt/mut) mice. P2Y12-CreER-GFP mice were crossed with DTA mice to obtain P2Y12-CreER-GFP::DTA (P2Y12+/CreER-GFP::DTAwt/mut) mice. CX3CR1GFP/GFP mice were crossed with C57BL/6 J mice to obtain CX3CR1+/GFP mice. P2Y12-CreER-GFP (P2Y12CreER-GFP/CreER-GFP) mice were crossed with C57BL/6 J mice to obtain P2Y12+/CreER-GFP mice. Mice of both sexes were utilized for the experiments unless specified. All the animals were housed in the specific pathogen-free (SPF) facility at the Animal Facility of Shenzhen Institute of Advanced Technology at Chinese Academy of Sciences, Department of Laboratory Animal Science at Fudan University and Animal Facility of Shanghai Mental Health Center at Shanghai Jiao Tong University School of Medicine. The animals were exposed to a 12-hour light/12-hour dark cycle and provided food and water ad libitum. The ambient temperature was kept at 20 °C to 26 °C, and the humidity was maintained between 40 and 70%.
The chemical reagents were purchased from Sigma‒Aldrich, and the cell culture media were purchased from Invitrogen unless otherwise specified. The CSF1R inhibitor PLX5622 was purchased from SYSE Bio (Cat#: JP-2112). PLX5622 was formulated into the AIN-76A diet at 1.2 g of PLX5622 per kilogram of diet by SYSE Bio (Cat#: D20010801). The normal AIN-76A diet (control diet, CD) was purchased from SYSE Bio (Cat#: PD1001). Ovomucoid (Cat#: A003085) and penicillin/streptomycin/amphotericin B solution (P/S/A, Cat#: B540733) were procured from Sangon Biotech. Percoll (Cat#: 17-0891-02) was purchased from GE Healthcare. FcR blocking reagent (Cat#: 130-097-679) and conjugated monoclonal anti-ACSA-2 antibodies (Anti-ACSA-2 MicroBeads) (Cat#: 130-097-678) were obtained from Miltenyi Biotec. BODIPY™ 505/515 (BODIPY, Cat#: D3921) and pHrodo™ Red (Cat#: P36600) were purchased from Thermo Fisher Scientific. Recombinant complement C4b (Cat#: RPB305Mu01) was purchased from Cloud-Clone. GFAP ELISA kit (Cat#: ab233621) was procured from Abcam. Normal donkey serum and normal goat serum were purchased from Jackson ImmunoResearch. Hydrocortisone (NSC 10483, Cortisol) was purchased from Selleck (Cat#: S1696). Putrescine was obtained from Merck (Cat#: 51799). Insulin (human) was acquired from Selleck (Cat#: S6955). Recombinant Human FGF-basic (154 a.a.) was purchased from PeproTech (Cat#: 100-18B). Recombinant murine EGF was purchased from PeproTech (Cat#: 315-09). Prostaglandin (PG) F2α was procured from Yeasen (Cat#: 60811ES03). Recombinant complement 4b (C4b) (endotoxin removal) (Cat#: RPB305Mu01), complement component 4a (C4a) (endotoxin removal) (Cat#: RPA389Mu01), complement component 1, Q subcomponent A (C1qA) protein (endotoxin removal) (Cat#: RPD207Mu01) and complement component 3 (C3) protein (endotoxin removal) (Cat#: RPA861Mu01) were purchased from Cloud-Clone. Recombinant mouse complement component C2 protein CF (C2) was acquired from R&D Systems (Cat#: 6725-SE-010).
To pharmacologically eliminate brain microglia, the mice were administered a PLX5622-formulated AIN-76A diet at libitum as previously described. In contrast, the control mice were fed the normal AIN-76A diet (control diet, CD) ad libitum. To sparsely label microglia, a low dose of tamoxifen (Sigma, C8267) at the dose of 15 mg per kg of body weight for CX3CR1-CreER::Ai14 or 50 mg per kg of body weight for P2Y12-CreER-GFP::Ai14 dissolved in corn oil (Aladdin, C116025) was intraperitoneally administered once to each mouse.
siRNA oligonucleotides targeting mouse Rubcn (CTCAGAGTAACAGGACCTT) were synthesized by OBiO Technology (Shanghai). Astrocytes were transfected with 50 nM Rubcn siRNA or the scramble control using jetPRIME® in vitro DNA & siRNA transfection reagent (Polyplus, Cat#: 114-15) according to the manufacturer’s protocol. Seventy hours after transfection, astrocytes were collected for Western blot detection and functional assays.
To specifically target astrocytes, we used AAV.PHP.eB expression plasmids expressing the mCherry reporter under the control of a truncated Gfap promoter (GfaABC1D). The short Gfap promoter is followed by a woodchuck hepatitis virus posttranscriptional regulatory element (WPRE) and growth hormone polyadenylation signal (pA). The plasmids are flanked by AAV inverted terminal repeats. All AAV expression vectors were constructed, packaged and molecularly verified by OBiO Technology (Shanghai). The miR30-based shRNA AAV vector targeting mouse Rubcn (CTCAGAGTAACAGGACCTT) was used for in vivo Rubcn knockdown, as verified in our siRNA knockdown assay. The designed AAV plasmids for Rubcn shRNA and scramble were constructed as pAAV-GfaABC1D-mCherry-mir30 Rubcn shRNA-WPRE and pAAV-GfaABC1D-mCherry-scramble-WPRE, respectively. The AAV vectors were intravenously injected into the mouse tail using ultrafine needle insulin syringes (BD). The 2-month-old C57BL/6 J, CX3CR1+/GFP and RUBICON−/− mice received a single intravenous injection of AAV PHP.eB Gfap-mCherry at 2 × 1011 vg/kg. AAV PHP.eB Gfap-mCherry-Rubcn shRNA and Gfap-mCherry-scramble control were intravenously injected into 2-month-old C57BL/6 J mice. These mice were tested for functional performance and sacrificed after 32 days for sample collection.
The mice were deeply anesthetized with a mixture of ketamine hydrochloride (100 mg per kg of body weight) and xylazine (10 mg per kg of body weight) by intraperitoneal injection. The animals were then sequentially perfused with 0.9% saline and 4% paraformaldehyde (PFA) (Sigma, 441244) in 0.01 M PBS. The brains were carefully collected and then post-fixed in 4% PFA in 0.01 M PBS at 4 °C overnight.
The brains were dehydrated in 30% sucrose in 0.01 M PBS at 4 °C for 2–3 days. After being embedded in optimal cutting temperature compound (OCT, Tissue-Tek), the brains were frozen in liquid nitrogen and stored at −80 °C before sectioning. Tissues with the regions of interest were sectioned using a cryostat (Leica, CM1950) at a thickness of 30 μm.
The brain sections were first subjected to three 10~15-minute rinses with 0.01 M PBS. The samples were then blocked in 4% normal donkey serum (NDS, Jackson ImmunoResearch, 017-000-121) or normal goat serum (NGS, Jackson ImmunoResearch, 005-000-121) in 0.01 M PBS containing 0.3% Triton X-100 (Sigma‒Aldrich T8787) (PBST) at room temperature (RT) for 1 h. Subsequently, the samples were incubated with primary antibodies with 1% NDS or NGS in PBST at 4 °C overnight. After three rinses with PBST, the samples were reacted with fluorescent dye-conjugated secondary antibodies with 4’,6-diamidino-2-phenylindole (DAPI, 1:500, Sigma‒Aldrich, Cat#: D9542-10MG, Lot#: 118M4025V) in 1% NDS or NGS containing PBST at RT for 1.5 h. The samples were then well rinsed three times and carefully mounted with antifade mounting medium (Vectashield H-1000 or DAKO S3023). The primary antibodies used in this study included the following: goat anti-mCherry (Biorbyt, Cat#: Orb11618, Lot#: J2446; 1:500), rat anti-CD31 (BD Biosciences, Cat#: 550274, Lot#: 9259767; 1:10), rabbit anti-IBA1 (Wako, Cat#: 019-19741 Lot#: WDK2121; 1:500), goat anti-IBA1 (Abcam, Cat#: ab5076, Lot#: GR3187278-2; 1:500), chicken anti-GFP (Abcam, Cat#: ab13970, Lot#: GR236651-12; 1:1,000), rabbit anti-S100β (Abcam, Cat#: ab52642, Lot#: GR252937-5; 1:300), rabbit anti-GFAP (Abcam, Cat#: ab7260, Lot#: GR297722-2; 1:300), rabbit anti-α-SMA (Abcam, Cat#: ab124964, Lot#: GR181740-66; 1:300), mouse anti-GFAP (Sigma‒Aldrich, Cat#: G3893, Lot#: 105M4784 V; 1:200), goat anti-PDGFR-β (R&D systems, Cat#: AF385, Lot#: BIWO619121; 1:200), rabbit anti-LC3 (Cell Signaling Technologies, Cat#: 4108, Lot#: 3; 1:100), rat anti-LAMP1 (1D4B) (Santa Cruz, Cat#: SC-19992, Lot#: C2715; 1:50), rabbit anti-PDGFR-α (Cell Signaling Technologies, Cat#: 3174 s, Lot#: 7; 1:500), mouse anti-CC1 (Millipore, Cat#: 14-0661-82, Lot#: 3129980; 1:200), rat anti-mouse C4 (Abcam, Cat#: ab11863, Lot#: GR3315169-2; 1:100), goat anti-GFP (Abcam, Cat#: AB6673, Lot#: GR3373716-2; 1:1,000), chicken anti-mCherry (Abcam, Cat#: ab205402, Lot#: GR3271744-8; 1:500), chicken anti-NESTIN (Abcam, Cat#: ab134017, Lot#: GR3291127-1; 1:200), rabbit anti-RFP (Abcam, Cat#: ab62341, Lot#: GR3319727-1; 1:1,000), rabbit anti-NeuN (Abcam, Cat#: ab177487, Lot#: GR3275122-6; 1:500), rabbit anti-MBP (Abcam, Cat#: ab40390, Lot#: GR297609-1; 1:200), rabbit anti-LAMININ (Sigma, Cat#: L9393-100UL, Lot#: 087M4889 V; 1:250), rabbit anti-GFP (Invitrogen, Cat#: A-11122, Lot#: 2273763; 1:1000), goat anti-OLIG2 (R&D, Cat#: AF2418, Lot#: UPA0719061; 1:400) and rabbit anti-PDGFα (Cell Signaling, Cat#: 3164 S, Lot#: 02/2020-6; 1:500). Secondary antibodies conjugated to Alexa Fluor 488 (AF488), AF568 and AF647 were diluted 1:600 unless otherwise specified. These antibodies included the following: AF488 donkey anti-rabbit (Thermo Fisher, Cat#: A11008, Lot#: 1829924), AF488 donkey anti-rabbit (Jackson ImmunoResearch, Cat#: 711-545-152, Lot#: 146247), AF488 donkey anti-goat (Jackson ImmunoResearch, Cat#: 705-545-003, Lot#: 145270, 1:2,000), AF488 donkey anti-goat (Thermo Fisher, Cat#: A11055, 1687906), AF488 donkey anti-chicken (Jackson ImmunoResearch, Cat#: 703-545-155, Lot#: 126602 and 122188), AF568 donkey anti-rabbit (Thermo Fisher, Cat#: A10042, Lot#: 1964370), AF568 donkey anti-goat (Thermo Fisher, Cat#: A11057, Lot#: 1871957 and 1640316), AF568 donkey anti-rat (Abcam, Cat#: Ab175475, Lot#: GR142910-1), AF488 goat anti-rat (Abcam, Cat#: Ab150157, Lot#: GR3189353-1) and AF647 donkey anti-rat (Jackson ImmunoResearch, Cat#: 712-605-153, Lot#: 142891). The validation data of each antibody are listed on the websites of the corresponding manufacturers.
Primary microglial and astrocyte cultures were established from the brains of P1 ~ P3 neonatal C57BL/6 J mice. After careful removal of the meninges and blood vessels, the mouse cortex was minced using a surgical blade and dissociated by gentle mechanical disruption in 0.025% trypsin for 5 min. The cell suspension was then quenched by 100% FBS and filtered through a 70-μm strainer. Primary microglia of high purity were isolated and purified from the mouse cortex as previously described with minimal modifications. In brief, the cell/tissue-containing T75 flask was shaken at 180 rpm for 30 min on an orbital shaker. The microglia-containing supernatant was collected, spun down and cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS, 1% penicillin/streptomycin and 20% LADMAC-conditioned media (produced from the LADMAC cell line CRL-2420). The resulting cells were purified microglia and ready to use the next day. The purity of the primary microglia was systematically evaluated in our previous study. Primary astrocytes were isolated and purified from the mouse cortex as previously described. In brief, the cells were centrifuged and resuspended in Dulbecco’s modified Eagle’s medium (DMEM) F-12 supplemented with 10% FBS and 1% penicillin/streptomycin (complete medium). The cells were then plated in a T75 flask, and the culture medium was changed every 3 to 4 days. Seven days later, microglia and OPCs were removed by shaking the flasks for 6 h at 240 rpm. The cells were then shaken for 18 h, and the culture medium was changed every 6 h. The purity of astrocytes was examined based on the expression of GFAP, S100β, IBA1, PDGFR-β, CD31, α-SMA, PDGFR-α, CC1, TUJ1 and NESTIN. For the serum-free primary astrocyte culture, one day after the purified astrocytes were plated, the serum-containing medium was removed by gently washing astrocytes with PBS three times and washing them with serum-free medium. The astrocytes devoid of serum were then cultured in chemically defined medium (CDM, serum-free medium supplemented with defined chemicals and growth factors), which promotes astrocyte survival but not reactivity, as previously described with minor modifications. The CDM contains DMEM F-12 supplemented with 1.2 mg/mL NaHCO3, 50 nM hydrocortisone, 100 nM putrescine, 500 ng/mL prostaglandin F2a, 50 ug/mL insulin, 100 ng/mL fibroblast growth factor (FGF), 200 ng/mL epidermal growth factor (EGF) and 1% penicillin/streptomycin.
Microglial cell death was induced by repeated freeze‒thaw cycles (3X) as previously described. The dead microglia were labeled with 2 μg/mL pHrodo™ Red at 37 °C for 30 min and then washed 3-4 times with 0.01 M PBS. To test whether the cell debris can be engulfed by cultured cells (e.g., primary microglia and astrocytes), pHrodo-labeled debris was added to the cell cultures at a debris-to-cell ratio of 10:1. The non-engulfed debris was thoroughly washed away with 0.01 M PBS. The engulfment of debris was then analyzed using a Nikon A1 laser-scanning confocal microscope. To test the role of complement opsonization in the astrocytic phagocytosis of microglial debris in serum-containing culture medium, FBS was heated at 56 °C for 30 min to inactivate complement components as previously described. The culture medium containing complement-inactivated FBS (complete medium-based formulation) was supplemented with 5 μg/mL C1q, C2, C3, C4a or C4b to test whether these supplements could restore debris engulfment in astrocytes. To test the requirement of each complement component for the astrocytic phagocytosis of microglial debris in serum-free culture medium, we cultured astrocytes in CDM that lacked serum. The CDM was first supplemented with 2 μg/mL C1q, C2, C3, C4a or C4b. We then tested which complement-supplemented CDM could restore the astrocytic engulfment of microglial debris at 72 h. Alternatively, the cultured astrocytes were incubated with pre-opsonized microglial debris with 5 μg/mL C1q, C2, C3, C4a or C4b overnight at 4 °C, and the astrocytic phagocytosis of cell debris after 72 h of cell culture was then analyzed.
C57BL/6 J mice were administered the PLX5622-formulated AIN-76A diet or the normal AIN-76A diet. Two days after treatment, the mice were deeply anesthetized and perfused with cold 0.01 M PBS containing heparin. The mouse brains were then carefully harvested and minced on ice immediately. Subsequently, the minced brain was trypsinized in a shaker at 37 °C and 100 rpm for 20 min, followed by the ovomucoid (2 mg/mL) neutralization. Three milliliters of L15 culture medium containing 0.5% bovine serum albumin (BSA) was then added, and the mixture was pipetted up and down. The upper-layer cell suspension was then passed three times through a 100-μm cell strainer. The cell suspension was transferred into a 15-mL tube, spun down and resuspended in L15 culture medium containing 30% Percoll. The myelin debris in the upper layer was then removed by gradient centrifugation at 500 × g for 10 min. The cell pellets were resuspended and blocked in L15 medium containing 0.5% BSA and 10% FcR blocking reagent at 4 °C for 10 min. The resuspended cells were then incubated with 10% anti-ACSA-2 MicroBeads at 4 °C for 15 min. The cells were then rinsed and resuspended in L15 medium containing 0.5% BSA. Astrocytes were enriched by flowing the cell suspension through the LS column (Miltenyi, 130-042-401) attached to a QuadroMACS separator (Miltenyi, 130-091-051) and washing the column three times with 3 mL of L15 medium containing 0.5% BSA. Immediately, the column was removed from the separator, and the magnetically labeled cells were flushed out with 5 mL of L15 medium containing 0.5% BSA. To increase the purity of ACSA-2-positive cells, the magnetic separation process was repeated two more times. The purified ACSA-2-positive cells were translocated in TRIzol and rapidly frozen in liquid nitrogen for subsequent RNA sequencing and qPCR.
For the RNA-seq analysis of astrocytes, FACS was utilized for the harvesting of tdTomato+ cells from tamoxifen-administered ALDH1L1-CreER::Ai14 mice. In brief, adult ALDH1L1-CreER::Ai14 mice were deeply anesthetized with a mixture of ketamine hydrochloride (100 mg per kg of body weight) and xylazine (10 mg per kg of body weight) by intraperitoneal injection. The animals were then perfused with cold 1X PBS, and the brains were then dissociated immediately and cut into 1-mm3 pieces by mouse stainless steel brain matrices (RWD). The tissue pieces were subsequently transferred into a C tube containing 3 mL of papain (8 U/mL, Sangon Biotec, S501621) digestion buffer. The C tube was then attached onto a gentleMACS Octo Dissociator (Miltenyi) with the 37C_ABDK program. At the end of the program, the C tube was detached and briefly centrifuged at room temperature. Subsequently, 10 mL of ice-cold DBPS containing 0.5% BSA was added, and the mixture was then pipetted up and down with a 1-mL pipette until large tissue clumps were detached. The dissociated cells were then filtered through a 70-μm cell strainer (Falcon) and centrifuged at 300 × g and 4 °C for 10 min, and the supernatant was discarded. The cells were resuspended in 4 mL of 30% Percoll (Sigma‒Aldrich) and centrifuged at 700 × g and 4 °C for 10 min to remove debris. The cells were washed once with 3 mL ice-cold DBPS containing 0.5% BSA, and the dissociated cells were collected for subsequent cell sorting. Before loading the cells onto a MoFlo Astrios EQ Cell Sorter (Beckman Coulter), the cells were stained with pSIVA-FITC (Abcam), which labels dead cells. After removal of the doublets and cell debris by FSC/SSC, approximately 1.2 to 1.4 × 105 pSIVA-FITC- (488, 530/30) tdTomato+ (561, 585/42) cells were sorted for subsequent analysis by pSVIA-FITC. The cell sorter was controlled by Summit, and the FACS data were analyzed using FlowJo 10.4.
After astrocytes were harvested by FACS, total RNA was extracted using TRIzol. The RNA purity and quantification were evaluated with a NanoDrop 2000 instrument (Thermo). The RNA integrity was assessed with an Agilent 2100 instrument (Agilent Technologies). Libraries were then constructed using a TruSeq Stranded mRNA LT Sample Prep Kit (Illumina) according to the manufacturer’s instructions. Transcriptome sequencing and analysis were conducted by OE Biotech. The libraries were then sequenced by Illumina HiSeq X Ten with the 150 bp paired-end (150 PE) platform, and 43.50 M to 50.84 M raw reads were generated from each sample. The raw fastq data were processed by Trimmomatic, low-quality reads were removed to obtain clean reads, and 41.41 M to 48.53 M clean reads were obtained from each sample. The clean reads were mapped to the mm10 mouse genome using HISAT2. The FPKM of each gene was calculated with Cufflinks. The read counts of each gene were then obtained by HTSeq-count. Differential expression was analyzed using edgeR 3.28.1. The DEG thresholds were set to |fold change | >= 2 and FDR <= 0.01.
Total RNA from cultured cells or MACS-sorted ACSA-2-positive cells was extracted with TRIzol. cDNA was reverse transcribed from total RNA using the ReverTra Ace™ qPCR RT kit (FSQ-101, TOYOBO) according to the manufacturer’s instructions. Subsequently, a 10-μL reaction system was prepared for qPCR using FastStart Essential DNA Green Master Mix (Roche, 06402712001) with a LightCycler 96 real-time PCR system (Roche, 05815916001). The relative cDNA concentrations of target genes were normalized to Gapdh. The primers used in this study were synthesized by BGI (Shenzhen, China) and included the following: Gapdh-forward (TGAGGCCGGTGCTGAGTATG −3ʹ), Gapdh-reverse (TGGTTCACACCCATCACAAACA), Gfap-forward (CACCTACAGGAAATTGCTGGAGG), Gfap -reverse (CCACGATGTTCCTCTTGAGGTG), C4b-forward (GGAGAGTGGAACCTGTAGACAG), C4b-reverse (CACTCGAACACGAGTTGGCTTG).
The cultured astrocytes were lysed in cold RIPA buffer containing 1% protease inhibitor cocktail. The protein concentrations of the cleared lysates were determined using the bicinchoninic acid (BCA) assay. Twenty micrograms of total protein was loaded into and resolved by a 12% SDS‒PAGE gel and subsequently transferred to a 0.45-μm PVDF membrane using a wet transfer system (Bio-Rad). The membrane was blocked in 5% nonfat milk in TBST for 1 h and incubated with rabbit anti-RUBICON (D9F7) antibody (Cell Signaling Technologies, Cat#: 8465, Lot#: 2; 1:1,000) and mouse anti-ß-actin (Santa Cruz Biotechnology, Cat#: SC-47778, Lot#: A2317; 1:3,000) overnight at 4 °C. After three 5-minute TBST washes, the membranes were probed with goat anti-rabbit (HRP) (Dako, Cat#: P0448, Lot#: 20034870; 1:3,000). The washed membranes were developed with an ECL kit. The scanned images were processed using Adobe Photoshop CS4.
The cultured astrocytes were lysed in RIPA buffer with 1% protease inhibitor cocktail. The cell lysates were centrifuged at 19,000 g for 30 min at 4 °C. The protein concentration in the supernatants was measured using a BCA kit. The protein level of GFAP in cell lysates was analyzed against GFAP antibody with the GFAP ELISA kit (Abcam, Cat#: ab233621) according to the manufacturer’s instructions.
BODIPY staining was performed to detect intracellular neutral lipid accumulation in cultured astrocytes. The cell samples were fixed with 4% PFA in 0.01 M PBS for 20 min and rinsed three times with PBS. The fixed cells were then incubated with 2 μM BODIPY staining solution in 0.01 M PBS at 37 °C for 15 min and then washed three times with PBS. The stained samples were imaged with a laser-scanning confocal microscope.
Confocal images of fluorescent specimens were obtained with a Nikon A1 confocal microscope equipped with a 20 × 0.75 numeral aperture (NA) Plan Apo objective and a 60 × 1.49 NA oil immersion objective or a Carl Zeiss LSM 900 confocal microscope equipped with the laser module URGB (diode laser 405 nm; diode laser 488 nm; diode (SHG) laser 561 nm and diode laser 640 nm) and the Airyscan 2. Plan-Apochromat 20x (0.8 NA) objective. Confocal images were captured with a distance interval of 0.5 µm between z-sections. The xy view of the confocal images is presented as maximal projections of z stacks, and xz or yz slice views of the regions of interest were reconstructed to illustrate the protein colocalization. For cell culture, images were acquired by a single focal plane. All confocal images were captured and processed using Nikon NIS-Elements AR (v.4.6) or ZEN 3.0 (Carl Zeiss), whereas brightness, contrast and gamma correction were performed if necessary. Confocal images with Z-stacks were utilized for 3D reconstruction using Imaris 9.7 (Oxford Instruments) (debris engulfment) or ZEN 3.0 (Carl Zeiss) (whole-mount retina).
The assay of the BBB integrity using dextran was mainly performed as previously described with modifications. In brief, mice were first fed a PLX5622-formulated AIN-76A diet or a control AIN-76A diet for 2 or 4 days. Subsequently, 75 μL of 10-kDa dextran-Alexa Fluor 647 (2 mg/mL) was injected through the mouse tail vein. Three hours after the injection, the mice were deeply anesthetized with a mixture of ketamine hydrochloride (100 mg per kg of body weight) and xylazine (10 mg per kg of body weight) via intraperitoneal injection. Subsequently, the mice were perfused with 0.9% saline to remove the residual tracer confined within blood vessels and then subjected to 4% PFA perfusion. The brain and liver were then collected and fixed with 4% PFA containing 0.01 M PBS at 4 °C overnight.
The percentage of tdTomato-negative microglia engulfing tdTomato-positive microglia was quantified from at least 500 tdTomato-negative microglia (N = 5 mice for each group), as counted under the 40X objective through the eyepiece of a Nikon A1 confocal microscope. To carefully determine the presence of any tdTomato-positive microglia within tdTomato-negative microglia, the z-axis was adjusted back and forth to ensure that the fluorescent signals deep in the tissue were included. Representative images were captured at 60X. The in vivo astrocyte phagocytosis of microglial debris was determined as the presence of GFP+ or IBA1+ puncta/fragments within mCherry-labeled astrocytes from approximately 700 to 1000 mCherry-labeled astrocytes in 5 to 7 mice. The majority of these counts were acquired from direct observation using a 40X objective from the eyepiece of a Nikon A1 confocal microscope. The Z axis was adjusted back and forth to ensure that all signals deep in tissue sections were included during the direct observation. To quantify the phagocytosis of microglial debris by astrocytes in the retina and other cell types as stained with their markers or labeled by the transgene, we captured approximately 15 to 20 fields and analyzed the colocalization between microglial debris (CX3CR1-GFP or IBA1 staining) and other cell types (markers or Nes-GFP) using Nikon NIS-Elements AR software. The colocalization was quantitively presented as the Manders overlap coefficient (r). Similar quantification processes were applied to the astrocyte phagocytosis of microglial debris in the 5xFAD model (N = 5 mice) and C4 colocalization with microglial debris (N = 5 mice for in vivo experiments and N = 5 independent biological replicates for in vitro experiments). The in vitro astrocyte phagocytosis of microglial debris was determined by quantifying the mean fluorescent intensity (MFI) of pHrodo-labeled microglial debris within astrocytes from several independent ×20 imaging fields (N = 5 independent biological replicates). The MFI was calculated using ImageJ (v.1.53), and the fold change was normalized against the control. The in vivo GFAP intensity following PLX5622 administration was determined by measuring the MFI of GFAP signals from 3 to 5 independent fields per mouse (N = 5 mice). The size of LC3-positive puncta in astrocytes treated with microglial debris or starvation was measured from 81 puncta of 3 independent biological replicates using ImageJ. Note that only LC3 puncta that encircle microglial puncta were included in the microglial debris-treated group. The numbers of LAPosomes in astrocytes were counted from at least 7 TEM fields (N = 4 and 5 independent biological replicates for the starvation and microglial debris-treated groups, respectively). Only those puncta showing the single-membrane structure with a diameter larger than 1 μm were included as putatively LAPosomes. The MFI of BODIPY lipids was quantified from 9 to 10 independent fields (N = 3 independent biological replicates) using ImageJ.
The statistical analyses were performed using Prism 8.4.0 (GraphPad) or R 3.6.1 (R Foundation). The results are presented as mean ± standard deviation (SD) (bar plot) or median ± quartile (violin plot). A two-tailed independent t test was used to compare the differences between two groups, whereas one-way analysis of variance (ANOVA) with Holm‒Sidak’s multiple comparisons test (post hoc) was used for comparisons among multiple groups. Statistical significance was defined as p (or adjusted p) ≤ 0.05. The exclusion criteria for experimental data points were death or severe sickness of animals during the experimental period. No outliers were excluded in this study. The animals and cultured cells were stochastically grouped from each experimental treatment or treatment condition. The results from the animal studies and cell cultures were evaluated independently by two blinded experienced researchers. No statistical methods were used to predetermine the sample sizes, but our sample sizes are similar to those reported in our previous publications. Each experiment was repeated in at least two independent batches to avoid bias among a single batch. Several biological replicates were included in each independent batch. A reanalysis of RNA-seq data (accession code GEO ID: GSE108269) was conducted using R (3.6.3) packages, including edgeR (3.28.1), pheatmap (1.0.12), cowplot (1.0.0), EnhancedVolcano (1.4.0), org.Mm.eg.db (3.10.0), clusterProfiler (3.14.3) and VennDiagram (1.6.20).
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Supplementary Information Supplementary Data 1 Reporting Summary | true | true | true |
|
PMC9592639 | 36178514 | Mai Soliman,Heba Shehta Said,Mohammed El-Mowafy,Rasha Barwa | Novel PCR detection of CRISPR/Cas systems in Pseudomonas aeruginosa and its correlation with antibiotic resistance | 30-09-2022 | Pseudomonas aeruginosa,CRISPR/Cas system,Multiplex–PCR,Antibiotic resistance,Biofilm formation | Abstract CRISPR (clustered regularly interspaced short palindromic repeats)-Cas (CRISPR-associated proteins) systems are considered as acquired immune mechanisms in Gram-positive and Gram-negative bacteria and also in archaea. They provide resistance/immunity to attacking bacteriophages or mobile genetic elements as integrative conjugative elements (ICE) as well as plasmid transformation. As an opportunistic pathogen, Pseudomonas aeruginosa has been held responsible for serious infections especially in hospitalized and immunocompromised patients. Three subtypes of type I CRISPR system (I-C, I-E, & I-F1) have been detected in P. aeruginosa genomes. In this work, P. aeruginosa isolates were collected from different clinical sources, and the three CRISPR/Cas subtypes (I-C, I-E, & I-F1) were detected via singleplex and multiplex PCR techniques using novel universal primers that were designed specifically in this study. CRISPR subtypes I-C, I-E, and I-F1 were detected in 10, 9, and 13 isolates, respectively. Furthermore, antimicrobial susceptibility of CRISPR/Cas-positive and negative isolates to different antibiotics and the capacity of biofilm formation were detected using disc diffusion method and tissue culture plate method, respectively. There was a significant correlation between the presence/absence of CRISPR/Cas system and both antimicrobial susceptibility to some antibiotics and biofilm-forming capacity among P. aeruginosa clinical isolates. Key points • A novel multiplex–PCR for detection of CRISPR/Cas-positive strains of P. aeruginosa. • Understand the correlation between CRISPR/Cas systems and other characters of P. aeruginosa. • Correlation between antimicrobial susceptibility and CRISPR systems in P. aeruginosa. Supplementary Information The online version contains supplementary material available at 10.1007/s00253-022-12144-1. | Novel PCR detection of CRISPR/Cas systems in Pseudomonas aeruginosa and its correlation with antibiotic resistance
CRISPR (clustered regularly interspaced short palindromic repeats)-Cas (CRISPR-associated proteins) systems are considered as acquired immune mechanisms in Gram-positive and Gram-negative bacteria and also in archaea. They provide resistance/immunity to attacking bacteriophages or mobile genetic elements as integrative conjugative elements (ICE) as well as plasmid transformation. As an opportunistic pathogen, Pseudomonas aeruginosa has been held responsible for serious infections especially in hospitalized and immunocompromised patients. Three subtypes of type I CRISPR system (I-C, I-E, & I-F1) have been detected in P. aeruginosa genomes. In this work, P. aeruginosa isolates were collected from different clinical sources, and the three CRISPR/Cas subtypes (I-C, I-E, & I-F1) were detected via singleplex and multiplex PCR techniques using novel universal primers that were designed specifically in this study. CRISPR subtypes I-C, I-E, and I-F1 were detected in 10, 9, and 13 isolates, respectively. Furthermore, antimicrobial susceptibility of CRISPR/Cas-positive and negative isolates to different antibiotics and the capacity of biofilm formation were detected using disc diffusion method and tissue culture plate method, respectively. There was a significant correlation between the presence/absence of CRISPR/Cas system and both antimicrobial susceptibility to some antibiotics and biofilm-forming capacity among P. aeruginosa clinical isolates.
• A novel multiplex–PCR for detection of CRISPR/Cas-positive strains of P. aeruginosa. • Understand the correlation between CRISPR/Cas systems and other characters of P. aeruginosa. • Correlation between antimicrobial susceptibility and CRISPR systems in P. aeruginosa.
The online version contains supplementary material available at 10.1007/s00253-022-12144-1.
CRISPR (clustered regularly interspaced short palindromic repeats)/Cas (CRISPR-associated proteins) systems are currently the highlight of biology research. In 1987, a notable repetitive DNA sequence was identified in E. coli genome while searching for the genes implicated in phosphate metabolism (Ishino et al. 1987). Soon after, it was named CRISPR. Later on, similar sequences were reported in archaea and a wide variety of bacteria (Ishino et al. 2018). CRISPR/Cas system comprised a CRISPR array that includes short repeats interspaced by distinct DNA sequences called spacers. Such spacers are located among a group of short Cas genes that are responsible for the CRISPR immunity function (England et al. 2018). Therefore, CRISPR/Cas system is considered an acquired immunity mechanism that stores the memory of foreign DNA of invaders, in distinctive spacer sequences of the CRISPR arrays (Koonin and Makarova 2019). The defense mechanism against foreign DNA via the CRISPR/Cas system includes three distinctive steps: immunization (spacer acquisition), CRISPR RNAs (crRNAs) biogenesis (expression), and target interference step (Barrangou 2013). The first stage includes the insertion of protospacers, which are pieces of foreign DNA from invading bacteriophages or plasmids into the CRISPR array. Protospacers act as a stored memory in the bacterial cell for the defense against the same plasmid or virus when it attacks the cell again. During the second stage, these spacers that are interrupted by repeats are expressed as small guide crRNAs. Lastly, in the interference stage, Cas proteins help crRNAs to quarry and damage invading bacteriophages or plasmids (England et al. 2018; Van der Oost et al. 2009). CRISPR/Cas systems are categorized into two groups (Class I & Class II) according to variation in Cas protein structure and sequence divergence among the effector modules (Makarova et al. 2020). Both classes comprise 6 types. Class I includes 3 types (I, III, & IV) and utilizes a multi-subunit crRNA-effector complex that carries on the mechanistic function of CRISPR immunity (adaptation, expression, and target interference). On the other hand, Class II includes the types II, V, and VI that has a single subunit crRNA-effector module (Makarova et al. 2015). Three types (I, II, and V) target/damage foreign DNA only; type VI modulates RNA only; whereas type III targets both RNA and DNA (Wang et al. 2019). CRISPR systems are less common in bacteria than archaea as they were detected in approximately 50% and 87% of bacterial and archaeal genomes, respectively (Makarova et al. 2013). Type I CRISPR system is equally distributed in both archaea and bacteria with complete single unit loci, whereas type IV and type V CRISPR systems are less common in both of them (Makarova et al. 2015). Type III CRISPR system is more profuse in archaea than bacteria, which possess type II systems chiefly (Makarova et al. 2015). CRISPR/Cas systems are present in Gram-positive and Gram-negative bacteria (Louwen et al. 2014). Further classification of these systems into subtypes has been adopted in bacteria harboring such systems. For example, in S. pyogenes, two CRISPR subtypes: type I-C (Class I) and II-A (Class II) were detected (Nozawa et al. 2011). In bacteria, CRISPR/Cas system plays a substantial role by providing resistance against plasmid transformation and invading viruses like bacteriophages (Barrangou et al. 2007; Marraffini and Sontheimer 2008). CRISPR/Cas cassettes harbor various loci that can uptake spacers from invading bacteriophages and hence provide immunity against recurrent infection by the same virus. It also can obtain spacers from self-replicating plasmids carrying antibiotic resistance determinants, leading to cleavage of plasmids (Garneau et al. 2010). CRISPR/Cas systems were correlated with the expression of some virulence genes in bacteria. In Pseudomonas aeruginosa, for example, it was proven to be important in swarming motility used. Zone of inhibition and biofilm formation that are significant features of P. aeruginosa pathogenicity (Palmer and Whiteley 2011). P. aeruginosa is an opportunistic microbe and a dominant cause of serious infections particularly in hospitalized and immunocompromised patients (Eladawy et al. 2021). It has been held responsible for several infections in the lungs (pneumonia), implants, and wounds, in addition to hospital-acquired infections (Lima et al. 2017). Three subtypes of type I CRISPR/Cas system were identified in the genomes of P. aeruginosa strains (Fig. S1): I-C, I-F1, and I-E (Cady et al. 2011; van Belkum et al. 2015). It was documented that the subtype I-F1 has the main function as an immune system, providing immunity against numerous viruses including bacteriophages (Cady et al. 2012). The standard P. aeruginosa strain PAO1 does not have a CRISPR/Cas system but the PA14 strain has a type I-F1 CRISPR system (Jeukens et al. 2014). Interestingly, the genomes of P. aeruginosa, which contain CRISPR/Cas systems, were relatively smaller than those missing CRISPR/Cas systems (van Belkum et al. 2015). According to our knowledge, there is no established PCR-based method for detection and differentiation between the known subtypes of type I CRISPR/Cas system in P. aeruginosa. Therefore, our current study aims to establish a novel detection method for different subtypes of the CRISPR/Cas system in P. aeruginosa. Moreover, the correlation between the existence of CRISPR systems in the bacterial isolates and both the antimicrobial susceptibility to different antimicrobial agents and biofilm-forming capacity was assessed.
All clinical samples were collected under the approval of Research Ethics Committee (Faculty of Pharmacy, Mansoura University, Egypt) with the ethical code 2020–80. Participants in this work have signed informed consents. The collection of clinical isolates, from Mansoura hospitals, took place for 7 months (June–December 2020). Bacterial isolates were identified as P. aeruginosa employing adequate microbiological laboratory techniques (Collee et al. 1996; Mackie 2006) by streaking on cetrimide agar media and the resulting colonies were examined after Gram staining under a microscope. Other confirmatory identification tests included testing for their ability to grow at 42 °C and oxidase and catalase production, in addition to the detection of the characteristic sweetish odor and green pigment after cultivation on cetrimide agar plates. For long-term storage, purified isolates were cultured in Muller Hinton Broth (MHB) medium and then preserved in the MHB containing glycerol (20% v/v) at − 80 °C. In all experiments, PAO1 was employed as a reference strain of P. aeruginosa.
Genes’ sequences (25 sequences for each gene, Table S1) belonging to different CRISPR-Cas cassettes of reported subtypes of P. aeruginosa were recovered from GenBank and subjected to alignment via the multisequence alignment program (Clustal Omega) that is provided by EMBL-EBI website (Goujon et al. 2010; McWilliam et al. 2013; Sievers et al. 2011). For further analysis of aligned sequences and detection of conserved regions, the Jalview program was used (Waterhouse et al. 2009). Different primers were designed at different conserved regions of different genes after fulfilling the criteria of optimum GC% (40–60%) and minimal probabilities of hairpin and primer dimer formation. Later on, primers fulfilling the previous criteria were further subjected to bioinformatic analysis via blast against the whole genomes of P. aeruginosa species using the free blast tool provided by the National Center for Biotechnology Information. Only primers showing specific binding with P. aeruginosa genomes were selected for further PCR experiments. Accordingly, primer pairs were selected for the following genes in each subtype’s CRISPR/Cas cassette: cas5, cas7, and cas8 genes for subtypes I-C and I-F1 in addition to cas5, cas7, and cas11 genes for subtype I-E (Table 1). The alignment of the conserved regions for primer design of the previously mentioned target genes is shown in Fig. 1. Full alignments of the whole sequences of these genes are demonstrated in Fig. S2.
The whole genome extracts of different isolates of P. aeruginosa were extracted by colony PCR as described previously (Eladawy et al. 2021; Englen and Kelley 2000). Briefly, isolates were cultured in MHB overnight and then streaked on cetrimide agar plates. A single separate colony from each isolate was suspended in nuclease-free water (50 μl) and boiled for 10 min. The suspension of the heat-lysed cells was centrifuged, and the supernatant (containing DNA extract) was kept at − 80 °C till further investigation. Each of the three selected genes of each CRISPR/Cas subtype in P. aeruginosa was detected by conventional singleplex PCR in the collected isolates using the designed primer pairs (Table 1). Each PCR reaction mixture included the following: 10 μl of 2 × Dream Taq™ Green PCR Master Mix (Thermo Scientific, USA), 0.8 μl of forward/reverse primers each (10 μM), DNA extract (volume containing 50 ng), and finally the mixture volume was adjusted to 20 μl. A negative control tube was incorporated in all PCR experiments using DNA extract of PAO1 strain. PCR conditions comprised an initial step of denaturation (95 °C/5 min), then 35 cycles of denaturation (95 °C/30 s), annealing (specific temperature for pairs of primers as demonstrated in Table 1/30 s), and extension (72 °C/60 s). At last, the reaction was concluded by a final extension (72 °C/5 min). Successful amplification of target PCR products, according to the amplicons’ size in Table 1, was assured by agarose gel electrophoresis employing 1.5% gels containing ethidium bromide followed by visualization using a gel documentation system (Acculab). Multiplex PCR technique was implemented for the detection of the CRISPR/Cas subtypes of P. aeruginosa in a single reaction tube. For such purpose, 3 strategies were followed in our study. The 1st strategy: included the use of a mixture of primer pairs (Mix1) that are designed for amplification of cas5 gene in the three different subtypes; I-C (amplicon size: 283 bp), I-E (amplicon size: 582 bp), and I-F1 (amplicon size: 829 bp). The 2nd strategy: a mixture of primer pairs (Mix2) that are used for amplification of cas7 gene in the different three subtypes; I-C (amplicon size: 318 bp), I-E (amplicon size: 1073 bp), and I-F1 (amplicon size: 763 bp) were used. The 3rd strategy: primer pairs’ mixture (Mix3) for the amplification of cas8 gene (subtypes I-C and I-F1) and cas11 (subtype I-E) with expected amplicon sizes of 274, 842, and 518 bp for cas8 (subtype I-C), cas8 (subtype I-F1), and cas11 (subtype I-E), respectively. The efficiency of the three strategies was evaluated in randomly selected CRISPR/Cas-positive isolates (3 positive isolates from each subtype).
A representative isolate from each CRISPR/Cas subtype was selected for sequencing of the three target genes of each subtype in this study. For such step, targeted genes were amplified employing Phusion High-Fidelity DNA Polymerase (Thermo Scientific, Scientific Inc.) and utilizing the specified primers in Table 1. PCR reactions were prepared based on the manufacturer’s recommendations. Purification of the target amplicons was performed via a gel extraction kit (Qiagen, Hilden, Germany). After purification, amplicons were shipped to Sigma Scientific Service Technical Support (Cairo, Egypt) for sequencing using Applied Biosystems 3500 XL Genetic Analyzer employing the reverse primers designed for each target gene. The chromatograms of the partial DNA sequences of the target genes were visualized and analyzed using the FinchTV program.
Antimicrobial susceptibility pattern of the 32 CRISPR/Cas-positive P. aeruginosa isolates and 32 randomly selected CRISPR/Cas-negative isolates from the same sources was conducted using the Kirby-Bauer disc diffusion technique (Bauer 1966). Susceptibility to various groups of antimicrobial agents was examined; therefore, antibiotic discs (Bioanalyse®, Turkey): piperacillin (100 μg), piperacillin-tazobactam (100/10 μg), ceftazidime (30 μg), cefepime (30 μg), amikacin (30 μg), imipenem (10 μg), meropenem (10 μg), ciprofloxacin (5 μg), and levofloxacin (5 μg) were used. Zone of inhibition diameter was determined and elucidated in accordance with Clinical and Laboratory Standards Institute guidelines (CLSI 2021).
The capacity of biofilm formation among the 32 CRISPR/Cas-positive P. aeruginosa isolates and the randomly selected 32 CRISPR/Cas-negative isolates was examined in vitro using the tissue culture plate method under static conditions as previously described (Stepanović et al. 2000). In brief, isolates were cultured in tryptic soy broth complemented with 1% anhydrous glucose (TSBG) and incubated at 37 °C overnight. Using the TSBG medium, bacterial cultures were adjusted at 600 nm to 0.2. Then, the adjusted cultures were transferred to a 96-well microtitre plate (100 μl/well, 4 wells/isolate). Following overnight incubation at 37 °C, bacterial cultures (from every well) were gently withdrawn and the microtitre plate was rinsed 3 times with phosphate-buffered saline (PBS, 200 μl/well) to discard non-adherent cells. The plate was dried in air, followed by adding 150 μl absolute methanol for fixation of the adherent cells. After the aspiration of methanol, the adhered biofilm was stained using 1% crystal violet (150 μl/well) for 20 min. To remove excess dye, the plates were washed gently three times using distilled water then the plates were kept inverted in the air until dry. Approximately 150 μl of glacial acetic acid (33% V/V) was transferred to each well for solubilization of the stained biofilm. Then, OD (at 540 nm) was assessed by means of ELx808™ Absorbance Microplate Reader (BioTek Instruments Inc., Winooski, VT). Negative control of medium only was included in each experiment (Di Domenico et al. 2016; Eladawy et al. 2020; Perez et al. 2011). To evaluate the biofilm-forming capacity of P. aeruginosa isolates, the mean optical density (ODi) of the bacterial isolate was determined. A cutoff value (ODC) was estimated as three standard deviations over the negative control’s mean OD. Isolates were considered non-biofilm producers (ODi ≤ ODC), weak producers (ODC < ODi ≤ 2 ODC), moderate producers (2 ODC < ODi ≤ 4 ODC), and strong producers when (4 ODC < ODi) (Lima et al. 2017; Stepanović et al. 2000).
Comparison of frequencies of distribution of CRISPR/Cas system and its subtypes, antimicrobial susceptibility, and biofilm formation capacity among P. aeuginosa isolates were evaluated via chi-square test or Fisher’s exact test (P < 0.05). Data was assessed via SPSS software (version 20.0; SPSS, Chicago, IL, USA).
In our work, 122 isolates of P. aeruginosa were collected from Mansoura hospitals for 7 months (June–December 2020). P. aeruginosa isolates were recovered from diverse clinical resources including: urinary tract infections (60 isolates), respiratory tract infections (13 isolates), wound exudate (24 isolates), blood (15 isolates), and other sources (10 isolates) like contact lenses, burns, and vaginal swabs samples (Table S2).
Conventional singleplex PCR indicated that 32 isolates harbored CRISPR/Cas systems. CRISPR/Cas-positive isolates were named using a capital letter referring to the detected subtype (C, E, or F1) followed by “c” (for clinical isolate) and lastly the number of the isolate, whereas CRISPR/Cas-negative isolates were given the name Nc tracked with the number of the isolate as mentioned in Table S2. CRISPR/Cas subtype I-C was detected in 10 isolates (Fig. 2A), while 9 isolates contained the subtype I-E (Fig. 2B) and 13 isolates harbored the subtype I-F1 (Fig. 2C). Each of the three selected genes (cas5, cas7, and cas8) of CRISPR/Cas subtypes I-C and I-F1 was detected in a separate PCR reaction for each isolate as shown in Fig. 2A and 2C. The cas5, cas7, and Cas11 genes were detected in each isolate harboring the CRISPR/Cas subtype I-E as demonstrated in Fig. 2B. Multiplex PCR was evaluated for its efficiency to detect the CRISPR/Cas subtype of P. aeruginosa in a single reaction. This was performed for all CRISPR/Cas-positive isolates using the primer mixtures: Mix1, Mix2, and Mix3. All primer mixtures specifically detected the CRISPR/Cas subtype, as determined in singleplex PCR, and representative samples are shown in Fig. 3. In multiplex PCR, CRISPR/Cas-negative strain (PAO1) was also included in the multiplex PCR experiments and did not show any interference with any of the adopted primer mixtures (Fig. 3).
The obtained partial sequences of the amplified genes in the CRISPR/Cas subtypes I-C, I-E, and I-F1 were successfully identified as the corresponding expected genes (100% identity) using the BLASTN search, on the NCBI website, against P. aeruginosa genome. Accordingly, the obtained sequences were uploaded to DDBJ/EMBL/GenBank database. The partial sequences of cas5, cas7, and cas8 genes in the isolates Cc5, Cc9, and Cc4 (subtype I-C) were given the accession numbers LC685202, LC685203, and LC685204, respectively, while those of cas5, cas7, and cas8 genes in the isolates Fc4, Fc2, and Fc5 (subtype I-F1) got the accession numbers LC685205, LC685206, and LC685207, receptively. Finally, the partial sequences of cas5, cas7, and cas11 genes in the isolates Ec5, Ec8, and Ec2 (subtype I-E) were given the accession numbers LC685208, LC685209, and LC685210, respectively.
The highest percentage of antibiotic resistance among CRISPR/Cas-positive isolates was observed for ceftazidime (100%) followed by cefepime (96.9%). While, the lowest percentage of resistance was detected against amikacin (18.7%), ciprofloxacin, and levofloxacin (21.9% each). Furthermore, the highest percentage of antibiotic resistance among CRISPR/Cas-negative isolates was noticed with ceftazidime (84.4%) followed by piperacillin (78.1%). While the lowest percent was observed with piperacillin-tazobactam (18.7%) and amikacin (21.9%) as shown in Table 2 and Table S2. Statistical analysis indicated a significant correlation (P < 0.05) between the susceptibility of clinical isolates to piperacillin and the presence/lack of the CRISPR/Cas system (Table 2). Higher sensitivity to piperacillin was observed among CRISPR/Cas-positive isolates (75%) in comparison with CRISPR/Cas-negative ones (21.9%). Moreover, the use of tazobactam in combination with piperacillin did not affect the sensitivity among CRISPR/Cas-positive strains while it increased the sensitivity among CRISPR/Cas-negative isolates (81.3%). There is no significant correlation between CRISPR/Cas subtypes and susceptibility to different antimicrobials tested (Table 2).
CRISPR/Cas-positive P. aeruginosa clinical isolates were categorized according to biofilm formation capacity into strong producers (18.8%), moderate producers (40.6%), and weak producers (25%). A low percent of these isolates did not produce biofilm (15.6%). In contrast among CRISPR/Cas-negative isolates, the highest percent were weak and non-biofilm producers (37.5% each), the lowest percent of isolates were strong producers about 9.4%, and 15.6% of the isolates were moderate producers (Table 3 and Table S2). Statistical analysis indicated a significant correlation (P < 0.05) between biofilm-forming capacity and the existence of CRISPR systems in clinical isolates of P. aeruginosa (Table 3). Most of the CRISPR/Cas-positive strains were strong/moderate biofilm producers in contrast with CRISPR/Cas-negative isolates that were weak or non-biofilm producers. CRISPR/Cas subtypes did not have a significant correlation with biofilm formation. Biofilm-forming capacity was nearly the same in the three subtypes (Table 3).
Studies that detected CRISPR/Cas systems in isolates of P. aeruginosa utilized computational approaches such as CRISPRCasFinder to detect CRISPR subtypes in sequenced genomes via Illumina sequencing (Silveira et al. 2020; van Belkum et al. 2015; Wheatley and MacLean 2021). CRISPRCasFinder is a software program that can be used for the prediction of CRISPR arrays and their associated proteins (Couvin et al. 2018). To our knowledge, there are no specific primers that can exclusively detect CRISPR/Cas systems and their subtypes in P. aeruginosa via PCR. Therefore, we were interested in designing specific primers for the detection of known CRISPR systems in P. aeruginosa via traditional PCR rather than expensive techniques of complete genome sequencing. Not all strains of P. aeruginosa species harbor CRISPR systems (van Belkum et al. 2015). Type I system, to which the three common subtypes in P. aeruginosa belong, can be distinguished from other types by its signature gene: cas3 (Makarova et al. 2015). However, the selection of cas3 gene as a target for PCR detection of CRISPR/Cas-positive strains of P. aeruginosa will be hindered by its high similarity to other genes, which are not related to CRISPR/Cas system as helicases (Makarova et al. 2015, 2020). In P. aeruginosa, subtypes I-C, I-E, and I-F1 include 7, 8, and 7 cas genes respectively (Fig. S1) (Makarova et al. 2020). Type I belongs to class I, whose subtypes can be classified according to sequence similarity clustering of effector proteins (Fig. S1) (Makarova et al. 2015). Interestingly, there is a tremendous difference in those proteins even within each subtype (Makarova et al. 2015). Therefore, genes in the effector complex of each subtype were selected as a target for PCR detection and were investigated for the presence of conserved regions after the alignment of 25 sequences for each gene. Moreover, the selected primers in the conserved regions were blasted against P. aeruginosa genomes to exclude those showing no specific binding elsewhere in the genome. Finally, we adopted in this work primers in the conserved regions of the following genes: cas5, cas7, and cas8 genes in the subtypes I-C and I-F1, in addition to cas5, cas7, and cas11 genes in the subtype I-E. It should be mentioned that we could not design a primer pair that can work for the same effector protein in the three known subtypes of CRISPR systems of P. aeruginosa. The reason for this is that we could not find any conserved regions for the selection of primers in the aligned sequences of all the subtypes harboring cas5, cas7, and cas8. After the successful detection of the selected gene of each CRISPR/Cas subtype in the CRISPR/Cas-positive isolates via conventional single PCR (Fig. 2), we were motivated to identify CRISPR/Cas-positive strains and further detect its subtype in a single reaction via multiplex PCR. For such purpose, we kept in mind, initially when selecting final primer pairs for singleplex PCR detection of genes of each subtype, that the resulting amplicons would differ in size to the extent that they can be resolved efficiently by agarose gel electrophoresis when used in mixtures for multiplex PCR. Therefore, it is clearly obvious that any of the primers’ mixture (Mix1, Mix2, or Mix3) would precisely and specifically detect CRISPR/Cas-positive isolates and also its subtype in a single-step reaction and without interference from PAO1 strain (CRISPR/Cas-negative strain) (Fig. 3). CRISPR systems were detected in 26.2% of the collected clinical isolates. In positive CRISPR/Cas subtypes, the most predominant subtype was I-F1 (40.6%), then I-C (31.3%), and lastly I-E (28.1%). van Belkum et al. (2015) indicated that the subtypes I-E and I-F1 are the most common subtypes found in P. aeruginosa. However, in the last study, the authors identified the subtype I-C for the first time; therefore, it is expected that its rate of detection would increase in the proceeding studies. A summary of the detected subtypes in this work is illustrated in the schematic diagram of Fig. 4 An elevated level of resistance to β-lactam antibiotics was detected among P. aeruginosa isolates clinically recovered from diverse sources in Mansoura hospitals, Egypt. Carbapenems, fluoroquinolones, or aminoglycosides could be employed for the management of infections caused by β-lactam-resistant isolates. Our results have indicated higher resistance to piperacillin among CRISPR/Cas-negative strains compared to CRISPR/Cas-positive ones (Table 2). Moreover, the use of tazobactam in combination with piperacillin did not affect the sensitivity among CRISPR/Cas-positive strains while it increased the sensitivity among CRISPR/Cas-negative isolates indicating the production of serine β-lactamases among CRISPR/Cas-negative-resistant isolates that are capable of hydrolyzing piperacillin that could be acquired through horizontal gene transfer (HGT). P. aeruginosa could develop resistance to various groups of antimicrobial agents through HGT or modification/mutation of the target site. HGT has an essential role in microorganism evolution and spread of resistance genes against different antimicrobial agents and is a main cause of genome expansion. Mobile genetic elements in P. aeruginosa transmit a wealth of genes that have been implied in a range of attributes including virulence formation (Louwen et al. 2014), xenobiotics degeneration, and antimicrobial resistance (van Belkum et al. 2015). CRISPR systems were initially known for their role in phage defense through inhibition of lysogenic conversion that is a principal mechanism of HGT. Later on, CRISPR/Cas systems were recognized for targeting other mobile genetic elements including plasmids or transformed DNA or integrative conjugative elements (Silveira et al. 2020; van Belkum et al. 2015; Wheatley and MacLean 2021). A previous study has indicated that strains of P. aeruginosa harboring CRISPR systems have considerably smaller genome sizes along with lowered mobile sulphonamide resistance genes (Shehreen et al. 2019). Recent studies have indicated that strains carrying CRISPR/Cas systems are characterized by smaller genome sizes and elevated GC content indicating that they hamper the transfer/gain of mobile genetic elements. The majority of strains harboring CRISPR systems have spacers targeting ICE and the conserved machinery for conjugative transfer of ICE and plasmids. Moreover, genomes harboring CRISPR/Cas systems have lowered the abundance of ICE and prophages. Remarkably, spacers of CRISPR systems delineate phages and ICE incorporated into the genome of strains lacking these systems (Silveira et al. 2020; van Belkum et al. 2015; Wheatley and MacLean 2021). In summary, this study provides a novel/innovative method for the detection of CRISPR/Cas-positive strains of P. aeruginosa via PCR and in a single reaction tube. We hope that this strategy would help to easily identify CRISPR/Cas-positive strains of P. aeruginosa. We believe that such convenience in identification will contribute to a better understanding of the correlational studies between the existence of CRISPR systems in P. aeruginosa and other characters of such pathogens, e.g., virulence, antibiotic resistance, and adaptation to environmental stress.
Below is the link to the electronic supplementary material.Supplementary file1 (PDF 8911 KB) | true | true | true |
PMC9592723 | Yuan Dong,Yuejie Zhang,Yingmei Feng,Wei An | The protective roles of augmenter of liver regeneration in hepatocytes in the non-alcoholic fatty liver disease 10.3389/fphar.2022.928606 | 11-10-2022 | augmenter of liver regeneration,non-alcohol fatty liver disease,non-alcoholic steatohepatitis,mitochondrion (mitochondria),hepatic/liver cells | Non-alcoholic fatty liver disease (NAFLD) occurs in 25% of the global population and manifests as lipid deposition, hepatocyte injury, activation of Kupffer and stellate cells, and steatohepatitis. Predominantly expressed in hepatocytes, the augmenter of liver regeneration (ALR) is a key factor in liver regulation that can alleviate fatty liver disease and protect the liver from abnormal liver lipid metabolism. ALR has three isoforms (15-, 21-, and 23-kDa), amongst which 23-kDa ALR is the most extensively studied. The 23-kDa ALR isoform is a sulfhydryl oxidase that resides primarily in the mitochondrial intermembrane space (IMS), whereby it protects the liver against various types of injury. In this review, we describe the role of ALR in regulating hepatocytes in the context of NAFLD. We also discuss questions about ALR that remain to be explored in the future. In conclusion, ALR appears to be a promising therapeutic target for treating NAFLD. | The protective roles of augmenter of liver regeneration in hepatocytes in the non-alcoholic fatty liver disease 10.3389/fphar.2022.928606
Non-alcoholic fatty liver disease (NAFLD) occurs in 25% of the global population and manifests as lipid deposition, hepatocyte injury, activation of Kupffer and stellate cells, and steatohepatitis. Predominantly expressed in hepatocytes, the augmenter of liver regeneration (ALR) is a key factor in liver regulation that can alleviate fatty liver disease and protect the liver from abnormal liver lipid metabolism. ALR has three isoforms (15-, 21-, and 23-kDa), amongst which 23-kDa ALR is the most extensively studied. The 23-kDa ALR isoform is a sulfhydryl oxidase that resides primarily in the mitochondrial intermembrane space (IMS), whereby it protects the liver against various types of injury. In this review, we describe the role of ALR in regulating hepatocytes in the context of NAFLD. We also discuss questions about ALR that remain to be explored in the future. In conclusion, ALR appears to be a promising therapeutic target for treating NAFLD.
Non-alcoholic fatty liver disease (NAFLD) is a spectrum of diseases that initially manifest as non-alcoholic fatty liver (NAFL) or non-alcoholic steatohepatitis (NASH). NAFL is characterized as simple steatosis without histological evidence of hepatocyte injury or inflammation, whereas NASH occurs with the presence of hepatic inflammation and ballooning degeneration (Schwimmer et al., 2005). When further aggravated, NASH may progress toward liver fibrosis and subsequently drive progression to advanced stages, including cirrhosis, hepatic decompensation, and hepatocellular carcinoma (Kim et al., 2018; Huang et al., 2021a). The liver has a powerful regenerative ability in many vertebrates. The augmenter of liver regeneration (ALR) is one of the key factors contributing to liver growth and regeneration (Gupta and Venugopal, 2018). The specific stimulatory and protective effects of ALR against various injuries have caught the eye of the scientific community, including researchers investigating NAFLD (Nalesnik et al., 2017). Mitochondrial dysfunction is a major contributor to the development of NAFLD, and ALR has critical mitochondrial functions (Kumar et al., 2020). Herein, we focused on the main isoform, 23-kDa ALR, located in mitochondria. This isoform protects hepatocytes against NAFLD via the modulation of mitochondrial homeostasis and mitophagy, suppression of oxidative stress, and promotion of cell regeneration.
With economic growth and changing lifestyles, the prevalence of NAFLD has increased rapidly to become a global burden. NAFLD occurs in 25% of the general population, especially in developed countries (Younossi et al., 2016; Cotter and Rinella, 2020; Gallego-Duran et al., 2021). It is estimated that the annual incidence of hepatocellular carcinoma is between 0.5% and 2.6% in patients with NAFLD and cirrhosis (Huang et al., 2021a). Given the vast proportion of patients with NFALD and its complications worldwide, the economic burden is huge.
The pathogenesis of NAFLD is not fully understood. Currently, the “multiple- or continuous-hit” hypothesis is the most accepted explanation (Gupta and Venugopal, 2018). An unhealthy lifestyle and its associated metabolic disorders, as well as genetic factors, contribute to the progression of NAFLD. Nonetheless, NAFLD is considered to result from an imbalance of energy metabolism in the liver (Mezhibovsky et al., 2021). Thus far, several genes have been identified to be critically involved in NAFLD.
Pennacchio et al. and Kim et al. identified the PNPLA3 gene as the most important genetic factor related to NAFLD to date (Romeo et al., 2008; Moon et al., 2022). The gene variant PNPLA3(148M) is a major risk factor for fatty liver, as it promotes steatosis through comparative gene identification 58- (CGI-58-) dependent inhibition of adipose triglyceride lipase (ATGL) during the progression of liver disease. CGI-58 is a cofactor of ATGL that significantly enhances the triglyceride (TG) hydrolase activity. Wang et al. speculated that the accumulation of PNPLA3(148M) sequestered the function of CGI-58, thereby limiting its access to ATGL or other lipases (Powell et al., 2019).
A variant of MBOAT7, which incorporates arachidonic acid into phosphatidylinositol (PI) (Lee et al., 2012), is also associated with the entire spectrum of NAFLD (Luukkonen et al., 2016; Mancina et al., 2016). The MBOAT7 gene encodes lysophosphatidylinositol acyltransferase 1 (LPIAT1), which preferentially binds arachidonic acid to PI (Lee et al., 2008), which is a constituent of membrane phospholipids and a precursor of phosphoinositide. Tanaka et al. demonstrated that the depletion of LPIAT1 in cultured hepatic cells caused a high PI turnover, which continuously produced diacylglycerol, a substrate for TG synthesis. This directly caused TG accumulation and collagen deposition within hepatocytes. Ultimately, this novel lipogenesis pathway is involved in the progression of NAFLD and may be a therapeutic target for NAFLD treatment (Tanaka et al., 2021).
The TM6SF2 gene encodes a protein involved in regulating hepatic TG secretion. A glutamic acid to lysine substitution at amino acid position 167 of the TM6SF2 protein (E167K) disrupts the secretion of very low-density lipoprotein (VLDL). Deletion of TM6SF2 resulted in abnormal VLDL-TG secretion, which progressed to hepatic steatosis (Carlsson et al., 2020). The lipidation of VLDL is a two-step process, with phospholipids and polyunsaturated fatty acids as key players in the second stage; TM6SF2 may also be involved in the second step of lipidation (Luo et al., 2022). Luukkonen et al. reported reduced levels of liver polyunsaturated fatty acids, serum TG, and hepatic phosphorylcholine in patients carrying the TM6SF2(E167K) variant. Knockdown of TM6SF2 in Huh7 and HepG2 cell lines reduced the expression of diacylglycerol O-acyltransferase 1 and 2 (Martin et al., 2021), which are two key enzymes in TG synthesis. In conclusion, the function of TM6SF2 is vital for the lipidation of VLDL (Luukkonen et al., 2017; Luo et al., 2022).
Lipodystrophy syndromes are extremely rare disorders of body fat deficiency associated with potentially serious metabolic complications, including diabetes, hypertriglyceridemia, steatohepatitis, and NAFLD (Brown et al., 2016). Mutations in genes associated with lipodystrophy, such as the peroxisome proliferator-activated receptor-gamma (PPARγ), lamin A/C (LMNA), and hormone-sensitive lipase genes, are potential therapeutic targets for NAFLD (DiStefano and Gerhard, 2022). PPARγ is part of the nuclear receptor family of transcription factors consisting of PPARγ, PPARα, and PPARδ (Liss and Finck, 2017). PPARγ performs different functions in various cells of the liver. In hepatocytes, PPARγ mediates the expression of adipogenesis genes, such as AP2 and CD36, which induce an increased uptake of free fatty acid (FFA). Simultaneously, the accumulation of FFA promotes intracellular TG accumulation (Chui et al., 2005; Wu et al., 2010). In hepatic macrophages, Kupffer cells (KCs) and monocytes, PPARγ promotes the activation of activated macrophages (M2) while inhibiting the activation of classical macrophages (M1). This reduces the release of inflammatory cytokines, such as tumor necrosis factor-α (TNF-α) and monocyte chemoattractant protein 1, and growth factors such as transforming growth factor-β (TGF-β), leading to reduced inflammation and activation of hepatic stellate cells (HSCs), consequently attenuating fibrosis. PPARγ is also associated with the quiescent phenotype of HSCs, limiting HSC activation and subsequent fibrosis (Skat-Rordam et al., 2019). Mahdi et al. described a 42-year-old female with lipodystrophy and NAFLD due to a pathogenic gene variant LMNA(D300N) (Peng et al., 2020). A polymorphism in the promoter of this hormone-sensitive lipase gene was associated with hepatic steatosis, obesity, diabetes, and dyslipidemia. Hsiao et al. found that patients with NAFLD often had complex metabolic abnormalities. Notably, the coexistence of NAFLD and glucose intolerance was shown to have a synergistic effect on increasing the body mass index, serum insulin levels, and homeostatic model assessment of insulin resistance. Body mass index and fat-insulin resistance, but not the homeostatic model assessment of insulin resistance, are consistent indices of insulin resistance in NAFLD studies. Thus, fat-insulin resistance may have the greatest effect on the elevation of serum TG in a state of glucose intolerance (Hsiao et al., 2013).
With economic growth, the global prevalence of metabolic syndromes, such as obesity, diabetes, and dyslipidemia, increases annually (Iacob and Iacob, 2022). Excessive intake of fructose, refined carbohydrates, sugar-sweetened beverages, saturated fat, and animal protein was identified as a major factor in the development of NAFLD (Parry and Hodson, 2017). For example, regular fructose consumption can induce hepatic lipogenesis and endoplasmic stress, impair fatty acid oxidation, deplete beneficial bacteria in the gut, and cause liver inflammation resulting from the production of uric acid and gut-derived endotoxins (Vos and Lavine, 2013; Jones et al., 2019). The World Health Organization reported that the number of obese people in China was below 0.1 million in 1975 and rose to 43.2 million in 2014, accounting for 16.3% of global obesity. A high-fat and high-carbohydrate diet and unhealthy lifestyle are the main causes of overweight/obesity and impair insulin resistance, which is key to the physiopathology of hepatic steatosis. Moreover, obesity-related hyperlipidemia worsens lipid metabolism disorders and is the most distinct feature of NAFLD. Given the increasing rates of obesity, type 2 diabetes mellitus, and other metabolic syndromes, coupled with an aging population, the incidence of NAFLD is projected to increase dramatically over time (Marjot et al., 2020). Mitochondrial dysfunction is frequently related to the development of NAFLD. Indeed, structural and functional alterations of mitochondria significantly contribute to changes in cellular lipid metabolism and oxidant stress responses (Auger et al., 2015). Domínguez-Pérez et al. found that cholesterol overload in the mouse liver induced by a high-cholesterol diet led to cholesterol and TG accumulation within hepatocytes, particularly their mitochondria. Moreover, this overload induced remarkable transcriptomic changes, mainly associated with mitochondrial function and dynamics favoring oxidative stress and apoptosis resistance, which could promote transformation (Dominguez-Perez et al., 2019). Dysfunction of hepatocyte endoplasmic reticulum (ER) homeostasis and the disturbance of its interaction with mitochondria also play an important role in NAFLD pathophysiology. The ER uses the unfolded protein response pathway to maintain protein and lipid homeostasis whenever exposed to hyperlipidemia, insulin resistance, inflammation, drugs, or other disturbances (Flessa et al., 2022). In addition to hepatocytes, KCs and HSCs are associated with the occurrence of NAFLD and progression to NASH during different stages of the NAFLD spectrum (You et al., 2008; Leroux et al., 2012; Teratani et al., 2012). The KCs are tissue-resident cells capable of self-renewal and the maintenance of liver homeostasis. Under normal conditions, KCs tend to suppress inflammation by secreting cytokines such as interleukin 4 (IL-4), IL-10, and IL-13 (Dixon et al., 2013). Hepatocytes express a series of membrane and cytoplasmic pattern recognition receptors, such as Toll-like receptor-4 (TLR-4), all of which stimulate KC activation and trigger a phenotypic switch of macrophages from M2 to M1 (Wan et al., 2014). Activated KCs promote the release of various inflammatory chemokines, including IL-1β, TNF-α, and IL-6 (Musso et al., 2013). These cytokines further recruit large numbers of monocyte-derived infiltrating macrophages to the damaged area, worsening inflammation and hepatocyte injury (Tacke, 2017). Chronic inflammation sustains these inflammatory stimuli and induces HSCs to initiate fibrotic processes. Located in the space of Disse in the liver, HSCs are physically quiescent cells that function as a store of vitamin A. Following sustained inflammation, HSCs are activated by cytokines and free radicals released from the surrounding cells, such as hepatocytes, T cells, and KCs. Once activated, HSCs undergo a phenotypic switch from adipocyte-like quiescent cells to myogenic cells with increased protein expression of alpha-smooth muscle cell actin (α-SMA) and extracellular matrix (ECM) proteins (Dooley et al., 2000). Increased ECM synthesis and reduced ECM degradation lead to excessive collagen deposition and the progression of fibrosis in the liver. Collectively, the crosstalk among hepatic cells substantially contributes to the development of NAFLD (Figure 1). Thus, determining how to protect hepatic cells from death and control the propagation of inflammation is essential for further understanding the pathology of NAFLD.
ALR was first identified in 1975 in crude extracts of liver homogenates in weaning rats (LaBrecque and Pesch, 1975). Injection of the purified substance into mice with partial hepatectomy stimulated liver regeneration. Therefore, it was named hepatic regenerative stimulator substance. More recently, it was formally named “augmenter of liver regeneration” (ALR). ALR is widely distributed in the testis, liver, kidney, brain, and other tissues, with maximum expression in the testis and liver (Lisowsky, 1996). Inside the liver, ALR is predominately expressed in hepatocytes and, to a lesser extent, in stellate cells (Gupta and Venugopal, 2018). Regarding subcellular localization, ALR is expressed in the nucleus and cytosol, as well as mitochondria (Weiss et al., 2017). Deletion of ALR was lethal in a yeast system (Becher et al., 1999).
The human ALR gene (growth factor erv1-like gene, GFER) is located on chromosome 16 and consists of three exons and two introns (Hagiya et al., 1994; Lisowsky et al., 1995), comprising a 299-bp 5ʹ untranslated region, a 375-bp coding sequence, and 550-bp 3ʹ untranslated region (Hagiya et al., 1994). A cDNA clone was more than 1.5-kb in length, and GFER has a “TATA-less” promoter (Shanks et al., 1993). Thus far, three isoforms of human ALR have been identified. The human ALR protein yields bands at 15-, 21-, and 23-kDa under reducing conditions, corresponding to 36-, 38-, and 40-kDa under non-reducing conditions, respectively (Dayoub et al., 2011; Dayoub et al., 2013; Gandhi et al., 2015; Weiss et al., 2017; Ibrahim et al., 2018). The 15-kDa ALR is secreted from hepatocytes into the extracellular environment, whereby it displays anti-apoptotic and anti-oxidative properties as well as inflammation- and metabolism-modulating effects (Ibrahim and Weiss, 2019). The 23-kDa ALR is a sulfhydryl oxidase that resides primarily in the mitochondrial intermembrane space (IMS), whereby it exerts liver protection effect against various types of injury (Mordas and Tokatlidis, 2015; Nalesnik et al., 2017; Weng et al., 2017; Jiang et al., 2019). Studies of 21-kDa ALR are limited. In this review, we focused on 23-kDa ALR.
The promoter region of GFER contains sites that bind inducers and repressors that positively or negatively regulate the ALR expression. Inducers include specific protein 1 (SP1), forkhead box A2 (FOXA2), early growth response protein 1 (Egr-1), and hepatocyte nuclear factor 4α (HNF4α). Repressors include activator protein 1/activator protein 4 (AP1/AP4), CCAAT/enhancer binding proteins (C/EBPβ), and HNF4α (Ibrahim et al., 2018). Binding of HNF4α to the promoter region (−209 to −204 bp) reduces GFER expression, whereas binding to another site (+421 to +432 bp) induces GFER expression (Guo et al., 2008). The downstream inducing effect of HNF4α is diminished upon the activation of the small heterodimer partner protein (SHP) (Ibrahim et al., 2018). There are two inducing response elements within the GFER promoter region. An upstream antioxidant response element (ARE) located between −27 and −19 bp induces GFER expression upon binding nuclear factor erythroid 2-related factor 2 (Nrf2) when hepatocytes are exposed to oxidative stress (Dayoub et al., 2013). In addition, a downstream site binding the IL-6 response element binding protein (IL-6-RE-BP) can increase the activating effect of FOXA2 (Dayoub et al., 2010). The promoter structure and regulation of the ALR gene are shown in Figure 2.
Mutations of GFER lead to severe mitochondrial disease. Fonzo et al. identified a c.581G→A homozygous mutation in the C-terminus of ALR, which results in a p. R194H substitution in children with autosomal recessive myopathy. At the cellular level, this mutation leads to respiratory chain defects, such as abnormal mitochondrial morphology and unstable mtDNA (Di Fonzo et al., 2009). Daithanker et al. characterized the R194H mutation in the context of enzymological studies of human ALR. The R194H mutation affected the thermal stability of mitochondria, as well their flavin adenine dinucleotide (FAD) binding and sensitivity to protein hydrolysis (Daithankar et al., 2010). The yeast ortholog Erv1p, a key protein in the mitochondrial disulfide relay system, oxidizes the disulfide carrier mitochondrial import and assembly protein 40 (Mia40), which in turn transfers disulfide bonds to newly synthesized small cysteine proteins in the IMS. Erv1p is then re-oxidized to transfer its electrons to molecular oxygen through interactions with cytochrome C and cytochrome C oxidase, linking the disulfide relay system to respiratory chain activity. Erv1p depletion prevents the import of these essential proteins, leading to mtDNA aberrations and abnormal mitochondrial morphology. Rat and human ALR proteins act as sulfhydryl oxidase and may play a role similar to that of yeast Erv1p (Di Fonzo et al., 2009). The features of ALR variants are summarized in Table 1.
At the organelle level, 23-kDa ALR is mainly located in the IMS, whereby it regulates mitochondrial biogenesis and function. The FAD-dependent sulfhydryl oxidase activity of ALR allows it to enhance the oxidative phosphorylation capacity of mitochondria (Daithankar et al., 2009). Compelling evidence indicates that ALR inhibits apoptosis, promotes hepatocyte regeneration, and prohibits fibrotic progression in various murine models of NAFLD (Xiao et al., 2015; Xu et al., 2016; Weiss et al., 2017; Kumar et al., 2020; Wang et al., 2020). All these beneficial effects of ALR in hepatocytes are directly or indirectly related to its regulation of mitochondria. Below, we focus on the protective effects of ALR in hepatocytes in the aspects of cell death, regeneration, and anti-fibrosis in NAFLD.
Excessive accumulation of TG, cholesterol, and lipid deposition are considered the first steps to induce hepatocellular lipotoxicity, followed by oxidative stress, lipid peroxidation, mitochondrial dysfunction, and excessive reactive oxygen species production (Dewidar et al., 2020). Taking cholesterol accumulation as an example, the activation of the adenosine monophosphate-activated protein kinase (AMPK) signaling pathway improves insulin resistance and lipid accumulation (Hsiao et al., 2013). Wang et al. showed that cholesterol accumulation within hepatocytes can be regulated by ALR via the liver kinase B1- (LKB1-) AMPK-sterol regulatory element binding protein 2- (SREBP2-) low-density lipoprotein receptor pathway. LKB1 is an upstream activator of AMPK. Knockdown of ALR expression inhibits LKB1 phosphorylation, leading to AMPK inactivation and SREBP2 maturation/nuclear translocation. SREBP2 and low-density lipoprotein receptor actions are closely associated with cholesterol accumulation within hepatocytes. Thus, alterations in these events can lead to extensive cholesterol accumulation and the development of lipid metabolism disorders (Wang et al., 2020). Results from many research groups have illustrated that ALR inhibits apoptosis and helps overcome cell injury induced by CCl4, ethanol, and other toxic factors. Studies in liver cells show that the downregulation of ALR results in increased activation of caspase-3 and caspase-9, an increased ratio of Bax/Bcl-2 expression, and reduced ATP content (Francavilla et al., 2014; Zhang et al., 2014; Dong et al., 2021). Beyond its role in cell apoptosis, there has been considerable evidence indicating a role for ALR in reducing autophagy. For example, in an in vivo ethanol-induced acute liver injury mouse model, the downregulation of ALR attenuated hepatotoxicity by activating autophagy, and in vitro experiments in the HepG2 cell line showed that protection was mediated by the inactivation of the Akt/mTOR pathway (Liu et al., 2019).
From mechanistic insights, mitochondria play a central role in hepatocyte survival. The transfection of ALR into steatotic hepatocytes upregulates carnitine palmitoyl transferase 1 (CPT1) expression to enhance long-chain fatty acid transport into mitochondria for usage (Xiao et al., 2015). Dynamin-related protein 1 (Drp1) is one of the major pro-fission proteins to clear damaged mitochondrial debris and govern mitochondrial homeostasis (Jin et al., 2021). In a murine model of hepatic ischemic reperfusion injury, the Drp1 activity increased, which promoted mitochondrial fission. Binding with transcription factor Yin Yang-1 (YY1) with ALR prohibited YY1 nuclear translocation and transcriptional activation. As one of the target genes of YY1, UBA2 is a subunit of the SUMO-E1 enzyme and catalyzes Drp1 SUMOylation. By doing so, ALR attenuated mitochondrial fission and retained its function (Huang et al., 2021b). Mitofusin-2 (Mfn-2) is an essential GTPase-related mitochondrial dynamics protein. In the same murine model of hepatic ischemic reperfusion injury, ALR administration accelerated Parkin translocation for transcriptional activation of Mfn2, leading to enhanced mitophagy (Kong et al., 2022). Most soluble IMS proteins rely on a mitochondria-targeting sequence for import, and ALR participates in protein import and export by cooperating with Mia40 (Finger and Riemer, 2020). Mia40 is reduced during the process of disulfide bond formation, and ALR can re-oxidize reduced Mia40 to make it available for the next round of disulfide bond formation (Grumbt et al., 2007; Bien et al., 2010; Banci et al., 2011). This function of ALR ensures adequate protein folding during import and export to the IMS, which is necessary for functioning mitochondria. Recent progress in the field revealed that the coiled-coil-helix-coiled-coil-helix domain-containing 4 (CHCHD4) proteins, the evolutionarily conserved human homolog of yeast Mia40, control antioxidant responses and lipid homeostasis (Reinhardt et al., 2020). Hence, CHCHD4/Mia40 could be a novel target for NAFLD investigations. Gandhi et al. successfully developed mice with liver-specific depletion of ALR (ALR-L-KO), which showed that a lack of ALR accelerated the development of steatohepatitis and hepatocellular carcinoma (Gandhi et al., 2015). Two weeks after birth, the ALR-L-KO mice showed reduced mitochondrial respiratory function, increased oxidative stress, and extensive steatosis and apoptosis. Furthermore, ALR depletion resulted in decreased expression of genes involved in lipid metabolism, such as CPT1α, and ATP synthesis, such as ATP synthase subunit ATP5G1. This model provides a useful tool to investigate the pathogenesis of steatohepatitis and its complications and further showed that ALR is required for mitochondrial function and lipid homeostasis in the liver.
Peroxisome proliferator-activated receptor-alpha (PPAR-α), CPT1-α, peroxisomal membrane protein 70 (PMP70), and acyl-CoA oxidase 1 (ACOX1) are a series of antioxidant proteins, which are targeted by miR540 (Kumar et al., 2019). In ALR-deficient hepatocytes, the miR540 expression was increased and the expression levels of PPARα, PMP70, ACOX1, and CPT1α were decreased. In contrast, antioxidant N-acetylcysteine and recombinant ALR rescued anti-oxidative stress responses by suppressing miR-540 expression and lipid accumulation in ALR-deficient hepatocytes. In agreement with these results, the exogenous administration of recombinant ALR to ALR−/−KO mice inhibited miR-540 expression and steatosis (Kumar et al., 2019).
FFA can induce steatosis and lipotoxicity, which are correlated with the severity of NAFLD. Moreover, the involvement of ER stress in lipotoxicity has been reported (Malhi and Kaufman, 2011). Xu et al. investigated the role of endogenous and exogenous ALR for FFA-induced ER stress and lipotoxicity. When hepatocytes treated with ALR or expressing ALR were incubated with palmitic acid in vitro, caspase-3 activity and Bax protein expression were reduced, therefore reducing lipotoxicity. These results indicate that ALR exerted its lipid-lowering and anti-apoptotic actions by elevating the mitochondrial FFA transporter CPT1α, increasing toxic FFA β-oxidation in mitochondria, and decreasing the delivery of toxic FFA metabolites. In vivo, reduced mRNA levels of ALR and FOXA2 (a transcription factor inducing ALR expression) were found in mice fed a high-fat diet, human patients with steatosis, and NASH liver samples. These results demonstrate the role of ALR in reducing lipid deposition and increasing β-oxidation in patients with NASH (Xu et al., 2016). Xiao et al. further confirmed that the protective role of ALR against steatosis occurred via the inhibition of calcium transport from the ER to mitochondria, and the inhibition of ER stress by ALR was associated with an interrupted interaction between Bcl2 and the inositol 1,4,5-trisphosphate receptor (IP3R) (Xiao et al., 2018).
The removal of 75% of rat liver tissue led to increased ALR mRNA expression levels in hepatocytes after 12 h, but DNA synthesis in liver tissue reached a peak 24 h later (Francavilla et al., 2014). This suggests that ALR is a significant factor in the process of liver regeneration (Fausto, 1991; Gandhi et al., 1999; Polimeno et al., 2011). The use of MitoBloCK-6 to pharmacologically inhibit ALR reduced the proliferation of hepatocellular carcinoma cells, an effect that links ALR function to mitochondrial iron homeostasis (Kabiri et al., 2021). Silencing of ALR inhibited the proliferation and triggered the apoptosis of U266 human multiple myeloma cells (Zeng et al., 2017). Conversely, ALR overexpression in hepatic cells enhanced cell proliferation via the microRNA-26a/p-Akt/cyclin D1 pathway (Gupta et al., 2019a). Kupffer cells (KCs) play a protective role in liver regeneration (Selzner et al., 2003), and a relationship between KCs and ALR has been reported. Yang et al. suggested that the activation of KCs was another mechanism by which ALR stimulates hepatocyte proliferation because there are high-affinity receptors for ALR on hepatic KCs, and ALR can stimulate KC proliferation (Wang et al., 2006). Similarly, when hepatocytes were co-cultured with KCs, the levels of hepatocyte DNA and protein in the supernatant were significantly increased (Kinoshita et al., 2005). These events indicate that ALR can regulate KCs to secrete certain growth factors which promote hepatocyte proliferation. Acute response cytokines, such as IL-6 and TNF-α, are mainly released from KCs and are associated with hepatocyte proliferation (Olthoff, 2002). As described above, ALR binds to KCs via high-affinity receptors. The activation of KCs induces the release of various cytokines that trigger hepatocyte proliferation.
Therefore, there is no evidence for a direct anti-fibrotic effect of ALR on hepatocytes. Nevertheless, ALR inhibits fibrotic progression in the liver by suppressing hepatic stellate cell (HSC) activation in NAFLD. Among all inflammatory cytokines, TGF-β1 is the most potent stimulator of HSC activation (Mu et al., 2018; Xiang et al., 2020; Cheng et al., 2021). The binding of TGF-β1 to its receptors on HSCs results in the phosphorylation of several serine and threonine residues, which stimulate Smad2 and Smad3 kinase activation and the formation of the Smad2/Smad3/Smad4 complex. This complex is then translocated into the nucleus, whereby it transcriptionally activates the expression of fibrotic genes, including the mitogen-activated protein kinase 1 (MAPK), phosphoinositide 3-kinase (PI3K), nuclear factor-κB (NF-κB), NADPH oxidase, and connective tissue growth factor genes (Pei and Li, 2021). In LX-12 cells treated with TGF-β1, miR-181 was upregulated, further increasing TGF-β receptor II expression on HSCs to potentiate fibrotic pathways. However, when LX-12 cells were transfected with an ALR plasmid, the overexpression of ALR counteracted TGF-β-induced miR-181 and TGF-β receptor II expression (Gupta et al., 2019b). Likewise, in cultivated renal tubular cells, the addition of human recombinant ALR decreased the TGF-β receptor II expression and phosphorylation of Smad2 and NF-κB (Liao et al., 2014). Metalloproteinases, which play a dominant role in ECM degradation, are inhibited by tissue inhibitors of metalloproteinase (TIMP). Of the four TIMPs, only TIMP-1 and TIMP-2 are detected in liver tissue, with TIMP-1 expression being more pronounced than TIMP-2. In a rat hepatic fibrosis model induced by porcine serum injection, the administration of ALR plasmid decreased the expression levels of TIMP-1 mRNA and protein and was accompanied by reduced deposition of collagen I and collagen II in the liver (Li et al., 2005). Cell motility is ATP-consuming and mediated by microfilament assembly. In vitro, ALR knockdown by shRNA promoted mitochondrial fission and elongation, which led to enhanced ATP production for HSC migration. Moreover, the proportions of F-actin and G-actin were higher in ALR-deficient HSCs following shRNA transfection. Conversely, ALR overexpression slowed HSC migration by reducing energy supply and inhibiting mitochondrial fusion (Ai et al., 2018). The proposed mechanism by which ALR protects hepatocytes in NAFLD is summarized in Figure 3.
Nearly 50 years of research on ALR has consistently demonstrated its involvement in the spectrum of NAFLD. Despite this progress, how ALR expression is regulated in the context of NAFLD is not well defined. Moreover, most ALR studies focus on the liver. Thus, whether ALR in other organs, such as the kidney and brain, communicate with the liver to participate in NAFLD remains an open question. Additionally, although 21-kDa ALR is one of the main isoforms, its role is not fully understood. There are still mysteries surrounding ALR worth exploring in the future. In conclusion, ALR promotes mitochondrial homeostasis, protects hepatocyte survival and function, and suppresses macrophage and HSC activation. Collectively, these features make ALR a potential therapeutic target for the treatment of NAFLD. | true | true | true |
|
PMC9593098 | Min Xiao,Song Yang,An Zhou,Tongxin Li,Jingjing Liu,Yang Chen,Ya Luo,Chunfang Qian,Fuping Yang,Bo Tang,Chunhua Li,Na Su,Jing Li,Mingying Jiang,Shiming Yang,Hui Lin | MiR-27a-3p and miR-30b-5p inhibited-vitamin D receptor involved in the progression of tuberculosis | 11-10-2022 | tuberculosis,vitamin D receptor,microRNA,macrophages,differentiation | Background MicroRNAs (miRNAs) play a vital role in tuberculosis (TB). Vitamin D receptor (VDR), an miRNA target gene, and its ligand, vitamin D3 (VitD3), have been reported to exert protective effects against TB. However, whether miRNAs can affect the progression of TB by targeting VDR has not been reported. Materials and methods Research subjects were selected according to defined inclusion criteria. A clinical database of 360 samples was established, including the subjects’ demographic information, miRNA expression profiles and cellular experimental results. Two candidate miRNAs, miR-27a-3p, and miR-30b-5p, were identified by a high-throughput sequencing screen and validated by qRT–PCR assays. Univariate and multivariate statistical analyses were performed. VDR and NF-kB p65 protein levels were detected by Western blot assays. Proinflammatory cytokine expression levels were detected by enzyme-linked immunosorbent assay (ELISA). Luciferase assays and fluorescence-activated cell sorting (FACS) were further applied to elucidate the detailed mechanisms. Results Differential miRNA expression profiles were obtained, and miR-27a-3p and miR-30b-5p were highly expressed in patients with TB. These results showed that the two miRNAs were able to induce M1 macrophage differentiation and inhibit M2 macrophage differentiation. Further experiments showed that the two miRNAs decreased the VDR protein level and increased proinflammatory cytokine secretion by macrophages. Mechanistically, the miRNAs targeted the 3′ untranslated region (3′UTR) of the VDR mRNA and thereby downregulated VDR protein levels by post-transcriptional regulation. Then, due to the reduction in VDR protein levels, the NF-kB inflammatory cytokine signaling pathway was activated, thus promoting the progression of TB. Conclusion Our study not only identified differentially expressed miRNAs between the TB and control groups but also revealed that miR-27a-3p and miR-30b-5p regulate proinflammatory cytokine secretion and macrophage differentiation through VDR in macrophages. Thus, these two miRNAs influence the progression of TB. | MiR-27a-3p and miR-30b-5p inhibited-vitamin D receptor involved in the progression of tuberculosis
MicroRNAs (miRNAs) play a vital role in tuberculosis (TB). Vitamin D receptor (VDR), an miRNA target gene, and its ligand, vitamin D3 (VitD3), have been reported to exert protective effects against TB. However, whether miRNAs can affect the progression of TB by targeting VDR has not been reported.
Research subjects were selected according to defined inclusion criteria. A clinical database of 360 samples was established, including the subjects’ demographic information, miRNA expression profiles and cellular experimental results. Two candidate miRNAs, miR-27a-3p, and miR-30b-5p, were identified by a high-throughput sequencing screen and validated by qRT–PCR assays. Univariate and multivariate statistical analyses were performed. VDR and NF-kB p65 protein levels were detected by Western blot assays. Proinflammatory cytokine expression levels were detected by enzyme-linked immunosorbent assay (ELISA). Luciferase assays and fluorescence-activated cell sorting (FACS) were further applied to elucidate the detailed mechanisms.
Differential miRNA expression profiles were obtained, and miR-27a-3p and miR-30b-5p were highly expressed in patients with TB. These results showed that the two miRNAs were able to induce M1 macrophage differentiation and inhibit M2 macrophage differentiation. Further experiments showed that the two miRNAs decreased the VDR protein level and increased proinflammatory cytokine secretion by macrophages. Mechanistically, the miRNAs targeted the 3′ untranslated region (3′UTR) of the VDR mRNA and thereby downregulated VDR protein levels by post-transcriptional regulation. Then, due to the reduction in VDR protein levels, the NF-kB inflammatory cytokine signaling pathway was activated, thus promoting the progression of TB.
Our study not only identified differentially expressed miRNAs between the TB and control groups but also revealed that miR-27a-3p and miR-30b-5p regulate proinflammatory cytokine secretion and macrophage differentiation through VDR in macrophages. Thus, these two miRNAs influence the progression of TB.
Tuberculosis (TB) is one of the top 10 causes of death and the leading cause attributable to a single infectious agent (Mycobacterium tuberculosis), ranking above HIV/AIDS. Approximately 1.7 billion people have contracted and been infected with M. tuberculosis worldwide, and 1.7 million people die from TB each year (Daley, 2019; Sinha and Hochberg, 2019; World Health Organization [WHO], 2019). Mycobacterium tuberculosis, the pathogen that causes TB, is an intracellular parasitic bacterium that infects humans. M. tuberculosis can inhibit the host immune response and escape immune surveillance (Goldberg et al., 2014). It is difficult to completely remove M. tuberculosis from the host, and once a host becomes infected, M. tuberculosis mostly causes latent infection and has a symbiotic relationship with the host (Jagielski et al., 2016). There are many theories about the mechanism underlying the intracellular pathogenesis of M. tuberculosis. However, these theories have not yet led to effective anti-TB treatments (Pieters, 2008; Jiang et al., 2018). The diagnosis of TB also needs to be improved and supplemented (McNerney et al., 2012). For these reasons, further research and exploration of the diagnosis and treatment of TB are still urgently needed. MicroRNAs (MiRNAs) play vital roles in promoting the progression of many diseases (Bertoli et al., 2015; Rupaimoole and Slack, 2017; Dai et al., 2022; Ren et al., 2022). MiRNAs can cleave or repress the mRNAs of target genes through the RNA-induced silencing complex (RISC) (Bartel, 2004; Mohr and Mott, 2015). Studies have reported that miR-155 expression is enriched in active TB (Etna et al., 2018). MiR-29a, miR-21, miR-99b, miR-652, and miR-146 were identified as potential novel biomarkers of TB and could be used to predict responses to treatment (Barry et al., 2018). The high expression of certain miRNAs in TB suggests that some miRNAs are related to the progression of TB. Vitamin D3 (VitD3), a steroid hormone, is thought to exert anti-inflammatory effects and play a vital function in innate immunity against intracellular pathogens (Wallis and Zumla, 2016). Studies have reported that VitD3 plays an important role in innate immunity against TB (Bekele et al., 2018; Jimenez-Sousa et al., 2018; Ayelign et al., 2020). VitD3 exerts its biological effects by binding to the vitamin D receptor (VDR) complex (Heikkinen et al., 2011; Cui et al., 2018), which activates and regulates multiple cellular pathways (Liu et al., 2006; White, 2012). It has been reported that the VitD3 levels in patients with active TB are lower than those in control individuals (Nnoaham and Clarke, 2008). Studies have shown that VitD3 is a protective factor in TB (Baeke et al., 2010). Studies have also reported that VDR, a receptor of VitD3, plays a vital role in TB. A decrease in VDR protein levels causes defects in VDR signaling, which impairs immunity against TB (Selvaraj et al., 2009). Polymorphisms in VDR, such as ApaI, BsmI, FokI, and TaqI, might affect susceptibility to TB (Joshi et al., 2014; Wu et al., 2015). It has been reported that miR-1204 is able to target VDR in breast cancer, leading to a poor prognosis (Liu X. et al., 2018). MiR-125a can target VDR to promote the occurrence and progression of liver fibrosis (He et al., 2021). Therefore, we wanted to explore whether any particular miRNAs are involved in TB regulation and are likely to depress TB by targeting VDR. To determine whether miRNA expression and VDR protein levels are correlated in TB, we studied these factors in a TB group and a control group. Subjects came from two medical institutions in Chongqing, and 181 TB patients and 179 control individuals were selected according to the inclusion criteria. A case report form including gender, living situation, education, cigarette smoking, body mass index (BMI), hypertension, etc., was used to obtain demographic information from the subjects. Peripheral venous blood samples collected from subjects were used for high-throughput sequencing and qRT–PCR validation. After univariate analysis and data processing, logistic regression analysis was performed, and two TB-associated miRNAs were screened. Finally, the two miRNAs were studied both phenotypically and mechanistically in monocytes by cytological experiments (Figure 1).
Human subjects were recruited for this study mainly from Xinqiao Hospital and Chongqing Public Health Medical Center between June 1, 2019, and December 31, 2019. Demographic information, clinical information, and peripheral blood specimens were collected from 360 subjects. Approximately 181 subjects were recruited into the case group, and 179 subjects were recruited into the control group, excluding previous exposure to TB. The patients were selected according to the following criteria: (1) patient was older than 18 years of age, of either sex; (2) patient was Han Chinese with a family that had lived in Chongqing for more than two generations; (3) patient had poisoning symptoms such as sputum, hemoptysis and emaciation, fatigue, night sweating, or low fever; (4) patient had at least one radiographic examination (X-ray and CT) suggesting pulmonary TB lesions; (5) patient had positive results on TB diagnostic tests such as sputum smear, sputum culture or blood molecular biology; (6) patient had Swiss cheese lesions or granuloma formation by pathological biopsy; and (7) patient was diagnosed with secondary pulmonary TB based on the symptoms, radiographic examination, etiological and pathological findings, regardless of whether they affected a single lung or both lungs and regardless of the presence of extrapulmonary TB. Patients were excluded according to the following criteria: (1) patient had consanguineous parents; (2) patient had other severe diseases (cancer, immune deficiency disease, pulmonary abscess, etc.); or (3) patient had latent TB. The control individuals were selected according to the following inclusion and exclusion criteria: (1) Individual who was older than 18 years of age, of either sex was eligible; (2) Individual who was Han Chinese with a family that had lived in Chongqing for more than two generations was eligible; (3) Individual without active or latent TB identified by clinical manifestations, radiographic examination and Purified Protein Derivative (PPD) tests was eligible. (4) Who with a history of TB exposure was excluded; or (5) Other exclusion criteria were the same as the TB group. Additionally, we collected clinical samples of peripheral venous blood for experimental studies, and from these samples, we isolated peripheral blood mononuclear cells (PBMCs) and plasma following appropriate experimental methods for subsequent research. All the data are available, and our research has been approved by the Xinqiao Hospital Ethics Committee.
Peripheral venous blood was collected, treated with the anticoagulation agent ethylenediaminetetraacetic acid (EDTA), and centrifuged to isolate blood cells and plasma. PBMCs in blood were separated by using density gradient centrifugation (Böyum and Scand, 1968; Harris and Ukayiofo, 1969). Blood was added down the wall of a tube containing human lymphocyte separation medium (Dakewe Co., Beijing, China, 711101X), and the PBMCs in the blood were purified by density gradient centrifugation (800 × g, 30 min). Plasma and PBMC specimens were stored at −80°C and in liquid nitrogen, respectively. The PBMCs used for RNA sequencing and miRNA/mRNA qRT–PCR were obtained from both TB patients and control individuals. PBMCs used for other cytological experiments were obtained from volunteers in the lab. Plasma specimens were used to measure concentration of 1, 25 (OH)2 D3 by enzyme-linked immunosorbent assay (ELISA) assay.
Following the manufacturer’s protocol for RNAiso Plus reagent (Takara Bio, Osaka, Japan), we extracted total RNA from the PBMCs of control individuals and TB patients. To obtain the miRNA differential expression profiles between TB patients and control individuals, the PBMC samples of five TB patients and five control individuals were selected to match the age and sex of the subjects. After total RNA extraction from PBMCs, we reverse transcribed the RNA into cDNA with the PolyA RT–PCR method. The cDNA was amplified by qPCR and then purified by PAGE. The differential miRNA expression profile was obtained by using high-throughput sequencing technology (LC-bio Technologies Co., Ltd., Hangzhou, China).
Total RNA was extracted with RNAiso Plus reagent (TaKaRa Bio, T9109). PrimeScript RT reagent kit with gDNA eraser (TaKaRa Bio, RR047A), SYBR qRT–PCR (TaKaRa Bio R4130-03) and the MiR-X miRNA First-Strand Synthesis Kit were used to reverse transcribe cDNA from mRNA and miRNA, respectively. The cDNA was analyzed by a qPCR kit (TaKaRa Bio RR820A) and GoTaq qPCR Master Mix (Promega, A6001, Madison, WI, United States) with a QuantStudio 3 and ViiATM7 quantitative real-time PCR instrument (Applied Biosystems, Waltham, MA, United States, with technical support by BioWavelet Co., Ltd., Chongqing, China). Gene-specific primers, oligo-dTs and random primers for the reverse transcription of miRNAs were synthesized by GeneCopoeia (Guangzhou, China) and Sangon Biotech (Shanghai, China).
We purchased the THP-1 and U-937 cell lines from the Shanghai Institute for Biological Sciences (Shanghai, China). The HEK-293T cell line was ordered from the National Infrastructure of Cell Line Resource (Beijing, China). We genotyped all cell lines and tested them for mycoplasma contamination though Shanghai Biowing Applied Biotechnology Co., Ltd. (Shanghai, China) Primary PBMCs were purified from the peripheral venous blood of the subjects by density gradient centrifugation as described above. We used DMEM and RPMI 1640 medium (HyClone, Logan, UT, United States) to culture HEK-293T cells and monocytes (THP-1, U-937, and PBMCs), respectively. We supplemented the culture media with 10% FBS (HyClone, Logan, UT, United States), 0.1 mg/ml streptomycin and 100 U/ml penicillin (Beyotime, Beijing, China). We cultured cells in an incubator (Thermo Fisher, Waltham, MA, United States) at 37°C, 1 atm and 5% CO2.
The proinflammatory cytokine levels in the culture medium and concentration of 1, 25 (OH)2 D3 in plasma samples were measured by ELISA. The ELISA kits were purchased from 4A BIOTECH company (Beijing, China), and the three proinflammatory cytokines that were assessed were interleukin-1 beta (IL-1β) (CHE0001, 96t), interleukin-6 (IL-6) (CHE0009, 96t), and tumor necrosis factor alpha (TNF-α) (CHE0019, 96t). The minimum concentration that could be detected was 7 pg/ml for all three kits. Centrifugation was performed at 800 rpm/min for 5 min to remove cells and obtain culture medium. Peripheral venous blood was treated with the anticoagulant EDTA and centrifuged (1100 rpm, 15 min) to obtain plasma. Cell culture medium and plasma were detected by ELISA. The reference range of 1, 25 (OH)2 D3 ELISA kit (EHC9044) is 15–60 ng/ml. All the experiments were performed in accordance with protocols of reagent kits. All experiments were repeated at least two times in triplicate.
The VDR coding sequence (CDS) was subcloned into pCDNA3.1 (+) by Sangon Biotech. We purchased pmirGLO dual luciferase plasmids carrying the VDR 3′ untranslated region (3′UTR) from YouBio Biological Company (Changsha, China). We named these plasmids WT, ΔWT1, ΔWT2, ΔWT3, ΔWT4, ΔWT5, ΔWT6, ΔWT7, and ΔWT8 and the corresponding mutant plasmids MUT1, MUT2, ΔMUT4, ΔMUT6, and ΔMUT8. We synthesized the miRNA mimics and inhibitors by GeneCopoeia Company (Guangzhou, China) and RiboBio (Guangzhou, China). We transfected the pmirGLO plasmids, miRNA mimics, and inhibitors into cells with Lipofectamine 3000 (Thermo Fisher, Waltham, MA, United States) and analyzed the cells at 48 h post-transfection. MiRNA mimics with Sulfo-Cyanine5 (Cy-5) dye were transfected into monocytes, and the transfection efficiency was observed by fluorescence microscopy (OLYMPUS IX83, UIS2 optical system). All experiments were repeated at least three times.
We harvested and lysed cells with 5 × loading buffer (Beyotime, Beijing, China) after culture for 48 h. We fully lysed the cell lysates with a vortex mixer (Thermo Fisher, Waltham, MA, United States) at 3000 rpm. The protein samples were then incubated in a dry bath incubator (Thermo Fisher, Waltham, MA, United States) at 100°C for 10 min to denature the proteins. We subjected protein samples to SDS–PAGE for gel electrophoresis, transferred the gel to PVDF membranes (Millipore, Boston, MA, United States) for transfer electrophoresis, and then exposed them by ECL (Thermo Fisher, Waltham, MA, United States). We purchased primary antibodies against VDR (D2K6W) and GAPDH (D4C6R) from Cell Signaling Technology (CST, Boston, MA, United States). We obtained HRP-conjugated antibody (ZB-2301, ZB-2305) from ZSGB-Bio Company (Beijing, China). All experiments were repeated at least three times.
We transfected the mimics into cells with the Lipofectamine 3000 reagent. The cells were cultivated for 36 h, and then the transcription inhibitor actinomycin D (Act D) (CST, Boston, MA, United States) was added to the cell cultures. Subsequently, we measured VDR mRNA expression by qRT–PCR in the treatment and control groups at 0, 3, 6, and 9 h. Additionally, we measured the expression levels of miR-27a and miR-30b-5p to assess the transfection efficiency. All experiments were repeated at least three times.
Monocytes were differentiated into different macrophage subsets by stimulation with different cytokines. THP-1 and U-937 cells were induced to polarize into M0 macrophages after stimulation by phorbol 12-myristate 13-acetate (PMA) (100 nM) 24 h later. Then, they were polarized to M1 macrophages after stimulation with lipopolysaccharide (LPS) (100 nM) and interferon-gamma (IFN-γ) (20 nM) or polarized to M1 macrophages after stimulation with interleukin-4 (IL-4) (20 nM) and interleukin-13 (IL13) (20 nM) 48 h later. PBMCs were first induced to differentiate into M0 macrophages by stimulation with human colony stimulating factor (h-CSF) (20 nM) for 5 days. Then, they were polarized into M1 or M2 macrophages as described above. Macrophage polarization was confirmed by flow cytometric analysis. We used cell surface markers to identify M0 (CD11b and CD68), M1 (CD40, CD64, CD86, etc.) and M2 (CD163, CD206, CD180, etc.) macrophages. After polarization was completed, we washed the cells once with PBS, scraped them gently and transferred them into fluorescence-activated cell sorting (FACS) tubes. We used fluorochrome-tagged monoclonal antibodies (Dakewe Co., Beijing, China) to stain the cells. CD11b (Pacific Blue) was used to identify M0 macrophages, CD64 (FITC) was used to identify M1 macrophages, and CD206 (PE) and CD163 (CY 7) were used to identify M2 macrophages. After labeling, we washed and resuspended the cells in PBS at least two times, and we analyzed the cells with FlowJo v10.5.3 by a Gallios flow cytometer (Beckman Coulter, Pasadena, CA, United States). All experiments were repeated at least three times.
After reaching 50% confluence, HEK-293T cells were prepared for exogenous nucleic acid transfection. The pmirGLO-VDR-3′UTR plasmids (WT, ΔWT1, ΔWT2, ΔWT3, ΔWT4, ΔWT5, ΔWT6, ΔWT7, ΔWT8, MUT1, MUT2, ΔMUT4, ΔMUT6, and ΔMUT8) were transfected into HEK-293T cells, and either 200 nM miRNA mimics or 200 nM miRNA inhibitors were also transfected at the same time. Forty-eight hours later, we measured the luciferase activity. All transfection experiments we performed were repeated at least three times in triplicate, and the luciferase activity data were normalized.
Continuous variable data are displayed as the mean ± SD, and categorical variable data are displayed as percentages according to the case group and control group. Statistical analyses were performed with SPSS 26.0 software, and VDR and miRNA expression levels were analyzed by GraphPad Prism 8.0. Univariate difference analysis was applied to all factors between the case group and control group. Logistic regression analysis was applied to the association between miRNA expression and other factors in the 360 subjects. Logistic regression and Pearson’s rank correlation test were used to calculate the correlation coefficients. For comparisons, the Mann–Whitney U-test was used if no significantly different variances existed between the two groups. To calculate the p-value, the unpaired t-test analysis was performed.
We collected general information about the subjects, such as age, sex, residence, and education, and analyzed this information by using univariate statistics (Table 1). Based on this basic information, there were significant differences in sex and age between the case and control groups. The TB group had a greater proportion of males and younger ages than the control group. Patients from rural areas and with low incomes and lower education levels accounted for a larger proportion of the case group. In addition, habits such as smoking and drinking also differed significantly between the case and control groups. Smoking and drinking behaviors seemed to increase susceptibility to TB. Poor nutrition, as indicated by a lower BMI, was more common in the TB group. These results are consistent with those of a previous report (Panda et al., 2019). We also found that diabetes was more common in the TB group than in the control group. Altogether, these results indicated that TB is a disease associated with many univariate factors.
We tried to determine whether any miRNAs that might influence TB progression were expressed at higher levels in the TB group than in the control group. MiRNA expression was measured in two steps. First, we performed high-throughput sequencing to determine the differential miRNA expression profile between the two groups (GES207224). Second, qRT–PCR was used in a case–control study to verify the miRNAs that were identified as being highly expressed in the differential expression profiles. These studies revealed that 8 out of 47 differentially expressed miRNAs were upregulated and 15 were downregulated in the TB patient group compared with the control group (Supplementary Figures 1A–C). We verified the expression of these eight miRNAs in 179 control individuals and 181 TB patients, and the results indicated that miR-27a-3p and miR-30b-5p had significantly higher expression in the TB group (Figure 2).
To determine whether the associations of miR-27a-3p and miR-30b-5p with TB confounded the effects of the other univariate factors, we corrected the other univariate factors. These two miRNAs, together with previously identified univariate factors related to TB, were analyzed by binary logistic regression (Table 2) and corrected with Model 1, Model 2, and Model 3. Multivariate statistical analysis revealed that after adjusting for the univariate factors in Model 1, Model 2, and Model 3, the two miRNAs were still closely associated with and highly expressed in the TB group. The results indicated that miR-27a-3p and miR-30b-5p have a significant association with TB.
Proinflammatory cytokines (IL-1β, IL-6, and TNF-α) were reported to play crucial roles in promoting the progression of TB (Nair et al., 2009; Gong et al., 2019; Ravan et al., 2019). To determine whether miR-27a-3p and miR-30b-5p are related to TB progression by regulating the secretion of proinflammatory cytokines, we transfected miR-27a-3p and miR-30b-5p mimics and inhibitors into cells and measured proinflammatory cytokine secretion. The results indicated that IL-1β and IL-6 secretion significantly increased in monocytes overexpressing either miR-27a-3p or miR-30b-5p. When miR-27a-3p or miR-30b-5p was knocked down, IL-1β and IL-6 secretion by monocytes was significantly reduced (Figures 3A,B). Similarly, the secretion of TNF-α was significantly increased in monocytes overexpressing miR-27a-3p or miR-30b-5p. In contrast, TNF-α expression was significantly decreased when either miR-27a-3p or miR-30b-5p was knocked down (Figure 3C). The above results indicated that miR-27a-3p and miR-30b-5p can promote proinflammatory cytokine secretion by monocytes. Monocytes were induced to differentiate by using established protocols (Zajac et al., 2013; Genin et al., 2015; Taniguchi et al., 2015). Following these protocols, monocytes were successfully induced to polarize into M1 and M1 macrophages (Figure 3D). To explore how miR-27a-3p and miR-30b-5p regulated and induced the differentiation of monocytes into macrophages, miR-27a-3p and miR-30b-5p mimics were transfected into monocytes, which were then induced to polarize into either M1 or M2 macrophages by using the previously described protocols. The cell surface expression of CD (cluster of differentiation) markers was measured to determine the differentiation ratio to assess the relationship between these two miRNAs and macrophage differentiation. Forty-eight hours after transfection with the two miRNAs and the induction of differentiation, the numbers of M0 (CD11b), M1 (CD64), and M2 (CD163, CD206) macrophages in each group were detected with flow cytometry. Generally, compared with that in the control group, the proportion of M1 macrophages was significantly higher in the experimental group. Correspondingly, compared with that in the control group, the proportion of M1 macrophages was significantly lower in the experimental group (Figures 3E–G). The above results indicated that miR-27a-3p and miR-30b-5p can promote monocytes differentiation into M1 macrophages but inhibit differentiation into M1 macrophages.
We predicted the downstream target genes of these two miRNAs. ENCORI database prediction suggested that the VDR complex (VDR/RXR-β) might be one downstream target gene of miR-27a-3p and miR-30b-5p (Table 3). There were possible binding sites of miR-27a-3p and miR-30b-5p in VDR target genes (Table 4). Moreover, the expression of VDR measured by qRT–PCR and the level of VitD3 detected by ELISA were significantly reduced in the TB group compared with the control group (Supplementary Figures 2A–D), suggesting that the two factors were also significantly related to TB. To prove whether the two miRNAs could regulate the protein level of VDR, we overexpressed miR-27a-3p and miR-30b-5p in M1 macrophages. After transfection for 48 h, we observed high expression of these two miRNAs in the experimental group by qPCR (Figure 4A). Moreover, under microscopy, the transfected cells exhibited strong fluorescence signals corresponding to these two miRNAs (Figure 4E). The above results indicated the successful overexpression of the two miRNAs in monocytes. Then, we measured the VDR protein levels in the experimental and control groups and found that miR-27a-3p and miR-30b-5p downregulated the protein level of VDR in the experimental group (Figure 4C). To fully elaborate these results, we knocked down either miR-27a-3p or miR-30b-5p expression in M1 macrophages. MiRNA expression and protein levels were measured separately after transfection with inhibitors. A significant decrease in miRNA expression (Figure 4B) and a significant increase in VDR protein levels (Figure 4D) were observed. However, in non-activated monocytes, neither miRNA significantly altered VDR protein expression (Supplementary Figures 3A–E). These results indicated that miR-27a-3p and miR-30b-5p are significantly involved in downregulating the protein level of VDR in activated M1 macrophages.
As previously predicted, the 3′UTR of VDR mRNA contains one binding site for miR-27a-3p and seven binding sites for miR-30b-5p (Figure 5A). We synthesized a pmirGLO reporter plasmid carrying the VDR sequence as well as a plasmid carrying mutant VDR sequences (Figure 5C). When miR-27a-3p and miR-30b-5p were overexpressed in HEK-293T cells transfected with the WT plasmid, the relative luciferase activity levels were significantly reduced (Figures 5B,D,F). When miR-30b-5p and miR-27a-3p were overexpressed in HEK-293T cells transfected with the MUT1 and MUT2 plasmids, respectively, the relative luciferase activity levels reverted to the previous levels (Figures 5B,E,G). Furthermore, we transfected the truncated mutant plasmids (ΔWT1-8) into HEK-293T cells and overexpressed either miR-27a-3p or miR-30b-5p. Both miR-27a-3p and miR-30b-5p significantly decreased the relative luciferase activities in the cells transfected with the ΔWT4, ΔWT6, and ΔWT8 plasmids. When miR-27a-3p and miR-30b-5p were overexpressed in HEK-293T cells transfected with the corresponding mutant plasmids (ΔMUT4, ΔMUT6, and ΔMUT8), the relative luciferase activities reverted to the previous levels (Figures 5B,H–J). The experiments described above indicate that there are binding sites in the 3′UTR of VDR mRNA that can be effectively bound by miR-30b-5p and miR-27a-3p. Previous studies have suggested that miRNAs impact target genes and downregulate their expression by post-transcriptional regulation (Correia De Sousa et al., 2019). Act D can inhibit cell transcriptional activity. Therefore, we treated cells overexpressing the two miRNAs with act D (Supplementary Figures 4A–F). The VDR protein level was significantly decreased compared with that in the control group when miR-27a-3p or miR-30b-5p was overexpressed in M1 macrophages (Supplementary Figures 4G–I). These results indicated that these two miRNAs could regulate VDR protein expression at the mRNA level, namely, by reducing the stability of VDR mRNA.
Previous experiments showed that miR-27a-3p and miR-30b-5p can promote proinflammatory cytokine secretion. Based on previous experiments, the VDR plasmid was overexpressed in cells first, and the cells were then transfected with miR-27a-3p or miR-30b-5p. The results showed that the proinflammatory cytokine levels in the experimental group reverted to the previous levels after VDR overexpression (Figures 6A–C). These experiments suggest that miR-27a-3p or miR-30b-5p promote the expression of proinflammatory cytokines in activated M1 macrophages by inhibiting the expression of VDR. Similarly, based on previous experiments, we transfected either miR-27a-3p or miR-30b-5p into monocytes after VDR overexpression. The results showed that macrophage differentiation in the experimental group reverted to the previous levels after VDR overexpression (Figures 6D–F). Gating strategy for flow cytometry assay is shown in Supplementary Figure 5. These experiments suggest that miR-27a-3p and miR-30b-5p induced M1 macrophage differentiation and inhibited M2 macrophage differentiation by inhibiting VDR.
VDR expression was highly expressed in both M1 and M2 macrophages, and expression in the latter was higher than that in the former (Figure 7A). MiR-27a-3p and miR-30b-5p expression are highly expressed in M1 macrophages, but lowly expressed in M2 macrophages (Figures 7B,C). We also found that there was a protein–protein interaction between VDR and NF-kB p65 and that VDR significantly inhibited NF-kB p65 protein expression (Figures 7D,E). Complete strips images of western blot are shown in Supplementary Figures 6–8. This finding suggests that one of the ways in which miRNAs regulate proinflammatory cytokine secretion and macrophage differentiation is by decreasing VDR protein levels, which at the same time activates the NF-kB signaling pathway and finally promotes the progression of TB. Taken together, the above results indicate that miR-27a-3p and miR-30b-5p play vital roles in promoting the progression of TB. On the one hand, by targeting and downregulating VDR mRNA expression, these two miRNAs upregulate the NF-kB signaling pathway and lead to the increased secretion of proinflammatory cytokines, thus promoting the progression of TB. On the other hand, by inducing the differentiation of classically activated M1 macrophages, these two miRNAs mediate antibacterial defenses in patients with TB (Figure 7F).
MicroRNAs regulate biological processes by targeting genes, thereby influencing the development of diseases (Huang et al., 2021). MiRNAs are expected to become new biomarkers for diagnosis, therapy prognosis prediction and treatment in breast cancer (Bertoli et al., 2015). Studies have shown that miRNAs target VDR in breast carcinoma and liver fibrosis and then affect disease progression (Liu X. et al., 2018; He et al., 2021), but whether these miRNAs are differentially expressed in TB is still unclear. It has been reported that miRNAs such as miR-92a-3p and miR-155 are highly expressed in TB and have an impact on the occurrence of the disease (Wu et al., 2012; Wang et al., 2018), but much work still remains to be done to clarify the mechanisms underlying this phenomenon. Some studies have shown that miRNAs play a vital role in the regulation of inflammatory cytokines (Zumkehr et al., 2018; Ban et al., 2020; Zhang et al., 2021) and that increased proinflammatory cytokine secretion by PBMCs or macrophages stimulated by other known factors can promote wound healing (Xie et al., 2021). Therefore, we wanted to screen miRNAs and investigate whether they meaningfully affect disease progression through known inflammatory pathways in TB. To determine which miRNAs can modulate the pathogenesis of TB and how they exert this function, the screened miRNAs should meet two conditions: first, they must be differentially expressed in TB. Second, they should have some target gene with a close connection to TB. Macrophages have a variety of physiological functions, including tissue repair, osteoclasts, antigen and antibody uptake, phagocytosis, antimicrobial effects, and antigen presentation, among others (Chistiakov et al., 2018). Macrophages have plasticity and phenotypic heterogeneity, can be induced to differentiate into M1 classical and M2 alternative activation forms, and play important roles in promoting and inhibiting inflammation in chronic inflammatory diseases (Parisi et al., 2018). Generally, M1 macrophages appear in the earlier stage of inflammation and can kill and remove pathogens and damaged cells, while M2 macrophages promote the regeneration and homeostasis of damaged tissues in the later stage (Atri et al., 2018; Zhou et al., 2019). Therefore, remodeling macrophage polarization to regulate the inflammatory process is considered a new prospective approach to treat inflammatory diseases. Many determinant factors can stimulate and induce macrophage polarization and shape macrophage phenotype and function, including metabolism and microbial metabolites, cellular metabolites, damaged cells, activated lymphocytes, miRNAs, inflammatory cytokines, interferon regulator factors, and epigenetic factors. Therefore, we chose to detect the differential expression of miRNAs in macrophages for the study of TB. Vitamin D receptor pathway activation shows certain anti-inflammatory effects. Research shows that the activation of VDR signaling represses inflammation and transforms Kupffer cells from a proinflammatory to an anti-inflammatory state (Dong et al., 2020). Another study suggested that VDR signaling induces macrophage differentiation and skews myelofibrosis (Wakahashi et al., 2019). The NF-kB pathway is considered to play a vital role in regulating inflammation. A study showed that the activation of NF-kB promotes M1 macrophage polarization and proinflammatory cytokine secretion, thereby promoting inflammatory disease progression (Lv et al., 2020). Research shows that VDR blocks NF-kB activation by decreasing the levels of the IKK complex (Chen et al., 2013), which was confirmed in our study. Many miRNAs target VDR in a specific manner. Through bioinformatics prediction and validation, we identified VDR as the target gene of the two miRNAs in this study. We also found that VDR could reverse the progression of M1 macrophage polarization and proinflammatory cytokine secretion induced by miRNAs. Therefore, in our study, we first screened miRNAs differentially expressed in TB and then investigated the function and mechanism of the identified miRNAs. Through comparisons of TB patients and controls, high-throughput sequencing and qRT–PCR were used to screen miR-27a-3p and miR-30b-5p, which were expressed at significantly higher levels in the TB group. Western blot experiments confirmed that the two miRNAs significantly inhibited VDR protein expression in differentiated macrophages but not in undifferentiated macrophages. Further experiments showed that miR-27a-3p or miR-30b-5p could significantly downregulate VDR mRNA expression in activated M1 macrophages. The miR-27a-3p and miR-30b-5p binding sites in the 3′UTR of VDR mRNA were verified by dual-luciferase assays. More specifically, ELISA and flow cytometry showed that the two miRNAs could promote M1 macrophage polarization and proinflammatory cytokine secretion. Moreover, after VDR overexpression, macrophage differentiation and proinflammatory cytokine secretion were reversed to a certain extent. The above results suggest that miR-27a-3p and miR-30b-5p induce M1 macrophage differentiation and upregulate the expression of proinflammatory cytokines by targeting VDR. Moreover, we found that the roles of these miRNAs in TB were previously reported in two different studies (Xin et al., 2016; Liu F. et al., 2018). These two studies clearly illustrated the important regulatory roles of these two miRNAs in TB. However, the relationship of miR-27a-3p and miR-30b-5p with VDR protein levels was studied here for the first time. In conclusion, this study identified miR-27a-3p and miR-30b-5p in TB patients. Both miRNAs target VDR mRNA, promote proinflammatory cytokine expression, and induce the polarization of monocytes into M1 macrophages while inhibiting the polarization of monocytes into M2 macrophages. Therefore, we conclude that the high expression of these two miRNAs is credibly related to the progression of TB.
The datasets presented in this study are deposited in the NCBI GEO repository, accession number GSE207224, available at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE20 7244.
The studies involving human participants were reviewed and approved by Xinqiao Hospital Ethics Committee. The patients/participants provided their written informed consent to participate in this study.
All authors listed have made a substantial, direct, and intellectual contribution to the work, and approved it for publication. | true | true | true |
|
PMC9593222 | 35859315 | Yuta Yoshino,Hiroshi Kumon,Tetsuya Shimokawa,Hajime Yano,Shinichiro Ochi,Yu Funahashi,Jun-ichi Iga,Seiji Matsuda,Junya Tanaka,Shu-ichi Ueno | Impact of Gestational Haloperidol Exposure on miR-137-3p and Nr3c1 mRNA Expression in Hippocampus of Offspring Mice | 21-07-2022 | Gene expression,RNA-sequencing,schizophrenia,haloperidol,gene ontology,prenatal exposure | Abstract Background Schizophrenia is a mental disorder caused by both environmental and genetic factors. Prenatal exposure to antipsychotics, an environmental factor for the fetal brain, induces apoptotic neurodegeneration and cognitive impairment of offspring similar to schizophrenia. The aim was to investigate molecular biological changes in the fetal hippocampus exposed to haloperidol (HAL) by RNA expression as a model of the disorder. Methods HAL (1 mg/kg/d) was administered to pregnant mice. Upregulated and downregulated gene expressions in the hippocampus of offspring were studied with RNA-sequencing and validated with the qPCR method, and micro-RNA (miR) regulating mRNA expressional changes was predicted by in silico analysis. An in vitro experiment was used to identify the miRNA using a dual-luciferase assay. Results There were significant gene expressional changes (1370 upregulated and 1260 downregulated genes) in the HAL group compared with the control group on RNA-sequencing analysis (P < .05 and q < 0.05). Of them, the increase of Nr3c1 mRNA expression was successfully validated, and in silico analysis predicted that microRNA-137-3p (miR-137-3p) possibly regulates that gene’s expression. The expression of miR-137-3p in the hippocampus of offspring was significantly decreased in the first generation, but it increased in the second generation. In vitro experiments with Neuro2a cells showed that miR-137-3p inversely regulated Nr3c1 mRNA expression, which was upregulated in the HAL group. Conclusions These findings will be key for understanding the impact of the molecular biological effects of antipsychotics on the fetal brain. | Impact of Gestational Haloperidol Exposure on miR-137-3p and Nr3c1 mRNA Expression in Hippocampus of Offspring Mice
Schizophrenia is a mental disorder caused by both environmental and genetic factors. Prenatal exposure to antipsychotics, an environmental factor for the fetal brain, induces apoptotic neurodegeneration and cognitive impairment of offspring similar to schizophrenia. The aim was to investigate molecular biological changes in the fetal hippocampus exposed to haloperidol (HAL) by RNA expression as a model of the disorder.
HAL (1 mg/kg/d) was administered to pregnant mice. Upregulated and downregulated gene expressions in the hippocampus of offspring were studied with RNA-sequencing and validated with the qPCR method, and micro-RNA (miR) regulating mRNA expressional changes was predicted by in silico analysis. An in vitro experiment was used to identify the miRNA using a dual-luciferase assay.
There were significant gene expressional changes (1370 upregulated and 1260 downregulated genes) in the HAL group compared with the control group on RNA-sequencing analysis (P < .05 and q < 0.05). Of them, the increase of Nr3c1 mRNA expression was successfully validated, and in silico analysis predicted that microRNA-137-3p (miR-137-3p) possibly regulates that gene’s expression. The expression of miR-137-3p in the hippocampus of offspring was significantly decreased in the first generation, but it increased in the second generation. In vitro experiments with Neuro2a cells showed that miR-137-3p inversely regulated Nr3c1 mRNA expression, which was upregulated in the HAL group.
These findings will be key for understanding the impact of the molecular biological effects of antipsychotics on the fetal brain.
Previous studies have shown that prenatal exposure to antipsychotics induces apoptotic neurodegeneration in the hippocampus and cognitive impairment of the offspring, which may then cause schizophrenia-like behavioral changes. Using pregnant mice exposed prenatally to haloperidol (HAL, 1 mg/kg/d), there were significant gene expressional changes, with 1370 upregulated and 1260 downregulated genes, compared with controls (P < .05 and q < 0.05) by RNA-seq of hippocampal RNA. GO analysis showed that synapse-related and neuron-related genes were prominent in both upregulated (563/1370 = 41.1%; >1.3-fold change) and downregulated (400/1260 = 31.7%; >1.3-fold change) genes. The expression of one microRNA, miR-137-3R, was increased in the HAL group, and it was confirmed that miR-137-3p regulates Nr3c1 mRNA expression with an in vitro luciferase assay.
Schizophrenia (SCZ) is a chronic mental disorder, and lifetime antipsychotic medication for it is essential to prevent relapse (Japanese Society of Neuropsychopharmacology, 2021). It is also recommended that women with SCZ continue antipsychotics even during pregnancy, because stopping medication often leads to a relapse of the psychiatric condition and because antipsychotics have lower teratogenicity than other medications (Hasan et al., 2015). However, the exact impact of exposure to antipsychotics on the neural development of offspring has not been well studied. Numerous studies have investigated the pathogenesis of SCZ based on several hypotheses, including dopamine (Seeman and Lee, 1975) and glutamate hypotheses (Javitt and Zukin, 1991). However, no clear mechanism has been elucidated. Both environmental and genetic backgrounds are associated with the onset of SCZ (Gomes et al., 2019; Socrates et al., 2021). MicroRNAs (miRNAs), which are noncoding RNAs approximately 20 nucleotides in length that are synthesized via several enzymatic processes, are an important component of the genetic background. A miRNA/miRNA* duplex forms, and 1 strand of the duplex loads into Argonaute homologue protein to form an RNA-induced silencing complex. Generally, such complexes work epigenetically as suppressors of mRNA expression (Yoshino and Dwivedi, 2020). MiR-137-3p was first identified in a genome-wide association study that reported the association between rs1625579 located in the miR-137 gene and SCZ onset (Schizophrenia Psychiatric Genome-Wide Association Study, 2011) and is a well-studied miRNA in SCZ research (Sakamoto and Crowley, 2018). Growing evidence has shown that miR-137-3p is associated with neurogenesis, synaptogenesis, and synaptic plasticity (Smrt et al., 2010; Szulwach et al., 2010; Siegert et al., 2015). It has also been consistently reported to regulate several genes relevant to neural functions (e.g., BDNF, CACNA1C, and TCF4) (Wright et al., 2013; Thomas et al., 2017). Prenatal stress, an environmental factor, is a risk factor for developing psychiatric disorders, including SCZ (Cattane et al., 2020; Krontira et al., 2020), and stress also regulates expression of several miRNAs in offspring (Liu et al., 2020a; Mazzelli et al., 2020; Labib, 2021). However, limited studies of miRNAs have been published. Among environmental background factors, maternal conditions (e.g., immunological changes and drug exposure) possibly contribute to SCZ pathogenesis (Glass et al., 2019; Choudhury and Lennox, 2021). Wang et al. (H. Wang et al., 2019) reported that prenatal exposure to antipsychotics (haloperidol [HAL] and risperidone) disrupts the plasticity of dentate neurons and memory. Prenatal exposure to risperidone also induces fetal neurotoxicity in the hippocampal region and cognitive impairment (Singh and Singh, 2017). Moreover, prenatal exposure to quetiapine impacts apoptotic neurodegeneration in the fetal hippocampus and cognitive impairment (Singh and Tripathi, 2015). The mechanism by which exposure of mothers to antipsychotics affects miRNA and mRNA expression in offspring is unknown. Thus, a prenatal HAL exposure mouse model was used to investigate (1) global expression changes by RNA-sequencing (RNA-seq) in the hippocampus of offspring; (2) the type of biological process based on differentially expressed genes (DEGs) using gene ontology (GO) analysis; (3) miRNAs relevant to global expressional changes according to in silico prediction; (4) the expression profiles of selected genes from RNA-seq, regulated by the miRNA; and (5) confirmation of the miRNA suppression of target genes using an in vitro system.
The outline of this study is shown in Figure 1.
Day 0 pregnant C57BL/6 mice (8 weeks old) were purchased from CLEA Japan (Tokyo, Japan). Pregnancy was confirmed by checking for a vaginal plug. The first generation of offspring was defined as the F1 model, and the pups born from F1 females and male mice were defined as the F2 model. Male mice (8 weeks) for mating with F1 females and a subchronic model (SC) aged 8 weeks were also purchased from CLEA Japan (Tokyo, Japan). The pregnant mice were housed 1 per cage (temperature, 22°C ± 2°C) with free access to water and food with a 12-hour light/-dark cycle (lights on at 6:00 am). All experiments were conducted according to the Guidelines for Animal Experimentation of Ehime University Graduate School of Medicine (Ehime, Japan) and the ethics committee of Ehime University Graduate School of Medicine (No. 28-34).
HAL and aripiprazole (ARP) were purchased from Tokyo Chemical Industry (Tokyo, Japan) and dissolved in 0.9% normal saline (NS) and 10% β-cyclodextrin (DEX) (Tokyo Chemical Industry), respectively. HAL and ARP were injected into pregnant females housed 1 per cage at 1 mg/kg/d (i.p., 2 times per day [7:00 am and 6:00 pm]), a dosing regimen capable of inducing side effects such as parkinsonism and catalepsy based on previous reports (Bilge and Erol, 2012; Nishchal et al., 2014; Sharma et al., 2018). Drug was administered from pregnancy day 1 to birth. Offspring (F1 and F2) were weaned at 4 postnatal weeks and housed 3 per cage after weaning. The SC mice were created by 1 mg/kg/d HAL or NS injection for 20 days from 8 weeks of age, and they were also housed 3 per cage. Only male mice were used for subsequent analyses. All mice were killed by decapitation, and tissue from the hippocampus was bilaterally dissected on an ice-cold stage, weighed, and stored at −80°C until analysis. This procedure was conducted according to the “Guidelines for the Care and Use of Mammals in Neuroscience and Behavioral Research (National Research Council (US) Committee on Guidelines for the Use of Animals in Neuroscience and Behavioral Research 2003).”
RNA was isolated using TRIzol (Thermo Fisher Scientific, Carlsbad, CA, USA) according to the manufacturer’s protocol (Roy et al., 2017). RNA quality and concentration were tested using the NanoDrop 1000 system (Thermo Fisher Scientific, Yokohama, Japan). RNA samples with a 260/280 ratio between 1.8 and 2.0 were considered pure.
RNA (1.0 µg) was used in 40-µL reaction mixtures to synthesize cDNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA). Reverse transcription (RT)-qPCR was used to measure mRNA expression levels by the StepOnePlus Real-Time PCR System (Applied Biosystems). The Predesigned qPCR Assay used Mm.PT.58.7811767 for Drd2, Mm.PT.58.12183825 for Htr2c, Mm.PT.58.41280327 for Gsk3b, Mm.PT.58.33547773 for Nrxn1, Hs.PT.58.42952901 for Nr3c1, and Mm.PT.39a.1 for Gapdh. RT-qPCR was conducted by using the PrimeTime Gene Expression Master Mix (Integrated DNA Technologies, Inc., Coralville, IA, USA) in a final volume of 10 µL. The mRNA expression levels were tested in duplicate, and outliers were removed according to the Outlier Calculator (https://miniwebtool.com/outlier-calculator/). The relative expression value was calculated using Livak’s ΔΔCt method (Livak and Schmittgen, 2001).
RNA (0.25–8.0 μg; 3.75 μL) was used in a 5-μL reaction mixture to synthesize cDNA for miRNA using the Mir-X miRNA First-Strand Synthesis Kit (Takara Bio Inc., Tokyo, Japan). Subsequently, the cDNA for miRNA was diluted 10 times. The miRNA expression levels were measured in duplicate using the Mir-X miRNA qRT-PCR TB Green Kit (Takara Bio Inc.) and StepOnePlus Real-Time PCR System (Applied Biosystems) with miRNA-specific primers (miR-137-3p: 5’-TTATTGCTTAAGAATACGCGTAG-3’) according to the manufacturer’s protocol. Briefly, the thermal cycling conditions were initial denaturation for 10 seconds at 95℃ and 35 subsequent cycles of denaturation for 5 seconds at 95℃, annealing and elongation for 20 seconds at 60℃, and a dissociation curve step (60 seconds at 95℃, 30 seconds at 55℃, and 30 seconds at 95℃). U6 included in the Mir-X miRNA qRT-PCR TB Green Kit was used as an internal standard. Outliers were removed according to the Outlier Calculator (https://miniwebtool.com/outlier-calculator/). The relative expression value was calculated using Livak’s ΔΔCt method (Livak and Schmittgen, 2001).
RNA from F1 mice (NS vs HAL; n = 6 each) was subjected to RNA-seq. The RNA integrity number was calculated by an Agilent 2100 Bioanalyzer (Agilent Technologies Inc., Santa Clara, CA, USA) and an Agilent RNA 6000 Nano Kit (Agilent Technologies Inc.) (supplementary Table 1). Subsequently, 200 ng total RNA from each sample was used for RNA-seq library preparation. An RNA-seq library of each sample was created using the Illumina TruSeq Stranded mRNA Sample Prep Kit (Illumina, Indianapolis, IN, USA) according to the manufacturer’s protocol. The quality of the average size (340–380 bp) was validated using an Agilent DNA1000 kit and Agilent 2100 Bioanalyzer (Agilent Technologies Inc.). The amount was determined using qPCR with the Kapa Library Quantification Kit (Illumina). The MiSeq Reagent kit V3 on a MiSeq system was used for sequencing (Illumina) based on pair-end reads (75 bp) according to the manufacturer’s instructions. The running cycle was set to 150 cycles.
Raw data files (FASTQ format) were extracted from the MiSeq system (Illumina). The reads were aligned with the reference genome (mm10) using TopHat software (Trapnell et al., 2009). Cufflinks was used for estimating expression levels (metric fragments per kilobase of transcript per million mapped reads [FPKM value]) of the known genes (Trapnell et al., 2010). The criterion for abundantly expressed genes per group was set as average FPKM values ≥1.0 in each group. Genes including an FPKM value of 0 in either the NS or F1 group were excluded. DEG analysis was performed using the edgeR package (Robinson et al., 2010). Significant DEGs were defined as P < .05 and q < 0.05. DEseq2 was used to create a volcano plot (Love et al., 2014). A heatmap based on significantly changed DEGs was generated using heatmap.2.
ClueGO plugin (Bindea et al., 2009) in the Cytoscape program (ver. 3.8.0) was used to perform GO of biological process (BP), molecular function (MF), and cellular component (CC). Statistical significance was set at P < .05 with the post-hoc Benjamini–Hochberg method. Functional connectivities of significant GO terms were generated to a graphical network using GO term fusion based on the following criteria: visual style = groups; GO term/pathway network connectivity = medium (kappa score = 0.50).
A module for the functional enrichment of target genes and their functional roles were considered as canonical and disease pathways using Fisher’s exact test (P value threshold set at <.05) by Ingenuity Pathway Analysis software (Qiagen, Valencia, CA, USA).
MiRNAs relevant to significant DEGs on RNA-seq were predicted by Ingenuity Pathway Analysis software (Qiagen) with the miRNA Target Filter function with the criteria of “experimentally determined” or “high confidence” downstream targets of the miRNAs. An in silico prediction algorithm for 8 prediction programs (miRWalk, Microt4, miRanda, miRDB, Pictar2, PITA, RNA22, and Targetscan under miRWalk version 2.0 software package) was used for target gene prediction. Next, genes that were present in >6 of them were defined as predicted target genes.
Neuro2a neuroblastoma cells were purchased from KAC (EC89121404-F0, Kyoto, Japan) and cultured in Dulbecco′s Modified Eagle′s Medium/Nutrient Mixture F-12 Ham (DMEM/F-12) containing 10% fetal bovine serum and penicillin and streptomycin (10 000 U/mL). The cells were incubated at 37°C in a 5% CO2 atmosphere, and the medium was replaced every 24 hours. Double-stranded RNA oligos (mmu-miR-137-3p mimic [C-310413-05-0002] and mmu-miR-137-3p hairpin inhibitor [IH-31043-07-0002]) were used (Horizon Discovery Group, Cambridge, UK). RNA oligos were transfected into Neuro2a cells using Lipofectamine RNAiMAX (Invitrogen, Grand Island, NY, USA) according to the manufacturer’s protocol and harvested 48 hours posttransfection for target gene expression analysis. This study was replicated in 2 independent batches.
The seed and mutant sequence of Nr3c1 gene was predicted using TargetScanMouse (http://www.targetscan.org/mmu_80/) and cloned into a pmirGLO vector (Promega Corporation, Madison, WI, USA). The pmirGLO was digested by PmeI (New England Biolabs, Ipswich, MA, USA) and XbaI (New England Biolabs) following the manufacturer’s protocol. The sequences of seed and mutant oligos are shown in supplementary Table 2, and they were annealed by annealing buffer (10 mM Tris-HCl pH 7.5, 1 mM ethylenediaminetetraacetic acid, and 10 mM MgCl2) with the following conditions: 68°C for 2 minutes, 37°C for 10 minutes, and incubated at room temperature for 5 minutes. Digested pmirGLO vector and annealed seed or mutant oligos were ligated with Ligation-Convenience Kit:Nippon Gene (Tokyo, Japan). The plasmid was subjected to transformation of the E. coli DH5α, and the sequence of the inserted DNA in the recombinant plasmid obtained was confirmed by Sanger sequencing with sequence primer as shown in supplementary Table 2. For the luciferase reporter assay, luciferase reporter vectors and mmu-miR-137-3p mimic were cotransfected into Neuro2a cells using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s protocol and harvested 48 hours posttransfection (n = 3 in each group).
Mice (8 weeks of age) were decapitated, and brains were removed and fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Specimens were dehydrated in a graded series of ethanol, infiltrated in xylene, and embedded in paraffin wax. The brain was processed routinely for paraffin sectioning. Sections were cut at 7 μm, dewaxed in xylene, and rehydrated with a graded series of ethanol. The sections were assayed using the terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) method using an apoptosis in situ detection kit (FUJIFILM Wako Pure Chemical Industries, Osaka, Japan).
Deparaffinized sections were incubated in a Phosphate-buffered saline (PBS) solution including 0.005% hydrogen peroxide for 20 minutes to suppress endogenous peroxidase activity. After a brief wash in PBS, the sections were exposed to 3% normal goat serum (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) and 5% bovine serum albumin in PBS and then rinsed in PBS. The sections were processed for immunohistochemistry, mainly with the caspase 3 antibody (Bioss antibodies, Woburn, MA, USA) as the primary antibody at a dilution of 5 µg/mL in PBS with 5% bovine serum albumin. The antigen-antibody complexes were visualized using a VECTASTAIN ABC and substrate kit (Vector Laboratories, Burlingame, CA, USA).
SPSS 22.0 software (IBM Japan, Tokyo, Japan) was used for the statistical analysis. Assessment of normal distribution was performed using the Shapiro–Wilk test. Average differences in mRNA and miRNA levels between 2 groups were assessed using the Student’s t test or the Mann–Whitney U test, and the average differences in 3 or more groups were assessed by 1-way ANOVA with the post-hoc Tukey test or the Kruskal-Wallis test with the post-hoc Steel-Dwass test. Statistical significance was defined at the 95% level (P = .05).
Drd2 mRNA expression was measured to elucidate the direct effect of HAL, because the main pharmacological effect of HAL is as an antagonist of the Drd2 receptor. The number of mice in each model was as follows: NS vs HAL (SC, 8 vs 8: F1, 8 vs 6: F2, 8 vs 8) and NS vs DEX vs ARP (F1, 8 vs 8 vs 5). Drd2 expression was significantly increased in the HAL group compared with the NS group in F1 (NS vs HAL: 1.0 ± 0.37 vs 1.47 ± 0.30, P = .035) but not in SC (NS vs HAL: 1.0 ± 0.54 vs 0.81 ± 0.12, P = .959) or F2 (NS vs HAL: 1.0 ± 0.27 vs 0.85 ± 0.18, P = .206) mice, as shown in supplementary Figure 1. In terms of ARP treatment, there were no significant changes (NS vs DEX vs ARP: 1.0 ± 0.10 vs 0.97 ± 0.08 vs 0.99 ± 0.08, P = .752; supplementary Figure 2A).
MiR-137-3p expression was significantly decreased in the HAL group compared with the NS group in SC (NS vs HAL: 1.0 ± 0.34 vs 0.64 ± 0.11, P = .022; Figure 2A) and F1 (NS vs HAL: 1.0 ± 0.13 vs 0.79 ± 0.09, P = .005; Figure 2B) mice. On the other hand, significantly increased miR-137-3p expression was found in the HAL group of F2 offspring (NS vs HAL: 1.0 ± 0.28 vs 1.30 ± 0.08, P = .029; Figure 2C).
A total 12 941 genes were identified using the stringent criteria described in the Methods; these genes are depicted in the volcano plot in supplementary Figure 3A. Of those genes, 1370 upregulated and 1260 downregulated genes were detected in the HAL group compared with the NS group (P < .05 and q < 0.05). A heatmap of significant DEGs is shown in supplementary Figure 3B. All significant DEGs are shown in supplementary Table 3.
A total 563 upregulated genes (>1.3-fold change) were subjected to analysis, and 87 BP terms reached significant levels (supplementary Table 4). Synapse and neuron-related BP terms were abundant and connected to each other (Figure 3A). Twelve CC and 14 MF terms reached significance. Interestingly, 4 synapse membrane-related CC terms were the most significant (e.g., postsynaptic density membrane, integral component of postsynaptic membrane, integral component of postsynaptic specialization membrane, integral component of postsynaptic density membrane). For the 400 downregulated genes (>1.3-fold change), 11 BP terms, 9 CC terms, and 1 MF term were significant (supplementary Table 5). When considering canonical pathways, the synaptogenesis signaling pathway was the most significant in terms of upregulated genes, but not in downregulated genes (Figure 3B). Several neurotransmitter receptor signaling genes (e.g., receptors for GABA, glutamine, dopamine, serotonin) were also significant among upregulated genes, but not downregulated genes. On the other hand, CNS-related terms such as neurological disease and psychological disease were found both among upregulated and downregulated genes (Figure 3C).
Based on the results of functional annotation and core analysis in the RNA-seq data, we focused on the 1370 upregulated genes because they were more relevant to neuronal functions than the 1260 downregulated genes. Subsequently, we narrowed down miR-137-3p from the list of miRNAs relevant to upregulated DEGs (supplementary Table 6) because miR-137-3p is well-studied in the SCZ field. Decreased miR-137-3p expression was found in the HAL group of F1 offspring. As a result of in silico prediction of miR-137-3p, 445 genes were considered to be predicted target genes based on our criteria (supplementary Table 7). Subsequently, 4 genes (Htr2c, Gsk3b, Nrxn1, and Nr3c1) were selected for qPCR validation according to the following criteria: (1) upregulated genes on RNA-seq; (2) predicted target genes of miR-137-3p; and (3) relevant to neuronal function or psychiatric disorders. qPCR in samples from F1 offspring validated Nr3c1 mRNA expression (NS vs HAL: 1.0 ± 0.13 vs 1.19 ± 0.14, P = .022; Figure 4A), but the other 3 genes were unchanged (Htr2c, P = .181; Gsk3b, P = .755; and Nrxn1, P = .182). Furthermore, decreased Nr3c1 expression was found in the HAL group of F2 offspring (1.0 ± 0.10 vs 0.87 ± 0.06, P = .016; Figure 4C) but not in SC mice (1.0 ± 0.10 vs 0.94 ± 0.10, P = .244; Figure 4B). In terms of ARP treatment, there were significant changes between the NS and DEX groups but not the DEX and ARP groups (NS vs DEX vs ARP: 1.0 ± 0.06 vs 1.11 ± 0.08 vs 1.03 ± 0.13, P = .045; supplementary Figure 2B).
When Neuro2a cells were transfected with the 20-nM miR-137-3p oligo, 40 times and 20 times upregulation of miR-137-3p expression was seen in the mimic compared with the vehicle in the first and second set, respectively (supplementary Figure 4). Nr3c1 and Gsk3b mRNA expressions were investigated because Htr2c and Nrxn1 mRNA expressions were not confirmed in the Neuro2a cell line. Significant downregulation was found for Nr3c1 expression in the first set (12%, P = .006; Figure 4D) and second set (22%, P = .047; Figure 4E). No significant changes were seen for Gsk3b in either the first (P = .059) or second set (P = .708).
As a result of in silico prediction, there were 2 possible seed sequences of mmu-miR-137-3p (AGCAAUA and GCAAUAA). The AGCAAUA sequence was selected for the dual luciferase reporter assay because the sequence is preserved in humans, whereas the GCAAUAA sequence is not (Figure 5A). When measuring the relative luciferase assay in pmirGLO vectors and 20-nM miR-137-3p co-transfected Neuro2a cells, there was a significant change (F = 6.188, P = .005; Figure 5B).
Differences were not observed in the TUNEL and caspase 3-positive reactions between sections from HAL-treated and control mice (supplementary Figure 5).
To the best of our knowledge, this is the first study to investigate miR-137-3p and global gene expression in offspring from mothers exposed to HAL. Based on the RNA-seq data, miR-137-3p, which is well-studied in the SCZ field, was selected from the list of miRNAs relevant to upregulated DEGs; miR-137-3p expression was downregulated and upregulated in F1 and F2 offspring, respectively. From the GO and core analyses with upregulated DEGs in F1 offspring, DEGs were relevant to synapse and neuronal functions. Of the DEGs, Nr3c1 mRNA expression was successfully validated in F1 offspring using the qPCR method, and downregulation of Nr3c1 mRNA was found in F2 offspring. Using an in vitro experiment, miR-137-3p was found to inversely regulate Nr3c1 mRNA expression by binding the 3’-UTR seed sequence of Nr3c1. Drd2 mRNA expression was upregulated in F1 offspring from mothers exposed to HAL. When considering the unchanged Drd2 expression in SC and F2 offspring, HAL affected fetuses by crossing the placenta. Indeed, placental passage of HAL was reported in a human study (Newport et al., 2007). Furthermore, the side effects of antipsychotics such as increased birth weight, abnormal neuromotor performance, and glucose tolerance were found in human offspring (Boden et al., 2012; Johnson et al., 2012; Frayne et al., 2018) and rodent studies (Courty et al., 2018). Interestingly, the opposite results for miR-137-3p expression (upregulation in AC and F1 offspring, and downregulation in F2 offspring) were found. Previous studies have also reported that HAL treatment induces expression changes in several miRNAs (Santarelli et al., 2013; Swathy and Banerjee, 2017; P. Wang et al., 2019). The same downregulation of miR-137-3p expression in AC and F1 offspring was directly affected by HAL treatment. In addition, a mouse study showed that administration of olanzapine to mothers induces decreased body weight in F1 offspring and increased body weight in F2 offspring (Courty et al., 2018). Taken together, opposite results at the molecular biological level between F1 and F2 offspring are understandable. From the RNA-seq results, upregulated DEGs were more relevant to synaptic and neural functions and several neurotransmitters than downregulated DEGs. Of those, we focused on Nr3c1, which encodes a glucocorticoid receptor, and its upregulated expression was successfully validated in F1 offspring. Elevated glucocorticoids in utero may induce neurodevelopmental alterations in offspring through the activity of the fetal hypothalamic–pituitary–adrenal axis (Krontira et al., 2020). Indeed, DNA methylation and mRNA expression changes of Nr3c1 have been reported in SCZ patients (Sinclair et al., 2012a; Sinclair et al., 2012b; Liu et al., 2020b). Other than downregulated Nr3c1, upregulation was found in 3 genes (Clock, Ncoa2, Ntrk2) on RNA-seq, which are also relevant to the glucocorticoid receptor signaling pathway (GO:0042921). The dysregulation of the glucocorticoid receptor signaling pathway is reported as an interaction with miR-137 (Valles et al., 2014). Using an in vitro experiment, miR-137-3p was found to inversely regulate Nr3c1 mRNA expression. Moreover, the luciferase activity assay showed that miR-137-3p regulated Nr3c1 mRNA expression through binding the predicted seed sequence in 3’-UTR of the Nr3c1 gene. Furthermore, inverse expression between miR-137-3p and Nr3c1 was found in F1 and F2 offspring. When considering the function of miRNAs as suppressors of mRNA expression, the miR-137 downregulation contributes to upregulation of Nr3c1 (and other targets) in the HAL group. Growing evidence has shown that miR-137-3p modulates neuronal apoptosis both in rats and in vitro (Wang et al., 2020) and apoptosis in hippocampal neural stem cells in mice (Schouten et al., 2015). In addition, apoptotic neurodegeneration was found in the hippocampus of rat offspring from mothers exposed to antipsychotics (Singh and Tripathi, 2015; Singh et al., 2016). Thus, we speculated that apoptotic changes occur in F1 offspring due to changes in the expression of miRNAs like miR-137-3p, but apoptotic changes were not found in our experiments. Apoptotic changes may need to be assessed in older mice because only 8-week-old offspring were used. This study has several limitations. First, animal studies cannot be applied directly to humans. In addition, the recommendations of pharmacotherapy for mothers with SCZ should not be affected by the present results. However, the present results indicate that offspring exposed to antipsychotics during gestation should be followed-up carefully in terms of not only teratogenicity but also other neurodevelopmental abnormalities. Second, behavioral tests were not conducted even though behavioral abnormalities, including cognitive dysfunction, were found in previous reports (Singh and Tripathi, 2015; Singh and Singh, 2017; Wang et al., H. 2019). In the future, replication studies using the same protocol should be conducted and include behavioral tests. In conclusion, it was confirmed that prenatal HAL exposure induces global gene expressional changes relevant to synaptic and neural functions. Of those, Nr3c1 mRNA expression was upregulated, followed by the downregulation of miR-137-3p, and this inverse expression was clarified in an in vitro experiment. These findings will be key to understanding the molecular biological mechanism of the effect of antipsychotics in the fetal brain.
Click here for additional data file. | true | true | true |
PMC9593258 | Paolo Petazzi,Laia Miquel‐Serra,Sergio Huertas,Cecilia González,Neus Boto,Eduardo Muñiz‐Diaz,Pablo Menéndez,Ana Sevilla,Núria Nogués | ABO gene editing for the conversion of blood type A to universal type O in Rhnull donor‐derived human‐induced pluripotent stem cells | 25-10-2022 | ABO,CRISPR,gene edition,iPSC,rare blood groups,Rhnull,transfusion medicine | Abstract The limited availability of red cells with extremely rare blood group phenotypes is one of the global challenges in transfusion medicine that has prompted the search for alternative self‐renewable pluripotent cell sources for the in vitro generation of red cells with rare blood group types. One such phenotype is the Rhnull, which lacks all the Rh antigens on the red cell membrane and represents one of the rarest blood types in the world with only a few active blood donors available worldwide. Rhnull red cells are critical for the transfusion of immunized patients carrying the same phenotype, besides its utility in the diagnosis of Rh alloimmunization when a high‐prevalence Rh specificity is suspected in a patient or a pregnant woman. In both scenarios, the potential use of human‐induced pluripotent stem cell (hiPSC)‐derived Rhnull red cells is also dependent on ABO compatibility. Here, we present a CRISPR/Cas9‐mediated ABO gene edition strategy for the conversion of blood type A to universal type O, which we have applied to an Rhnull donor‐derived hiPSC line, originally carrying blood group A. This work provides a paradigmatic example of an approach potentially applicable to other hiPSC lines derived from rare blood donors not carrying blood type O. | ABO gene editing for the conversion of blood type A to universal type O in Rhnull donor‐derived human‐induced pluripotent stem cells
The limited availability of red cells with extremely rare blood group phenotypes is one of the global challenges in transfusion medicine that has prompted the search for alternative self‐renewable pluripotent cell sources for the in vitro generation of red cells with rare blood group types. One such phenotype is the Rhnull, which lacks all the Rh antigens on the red cell membrane and represents one of the rarest blood types in the world with only a few active blood donors available worldwide. Rhnull red cells are critical for the transfusion of immunized patients carrying the same phenotype, besides its utility in the diagnosis of Rh alloimmunization when a high‐prevalence Rh specificity is suspected in a patient or a pregnant woman. In both scenarios, the potential use of human‐induced pluripotent stem cell (hiPSC)‐derived Rhnull red cells is also dependent on ABO compatibility. Here, we present a CRISPR/Cas9‐mediated ABO gene edition strategy for the conversion of blood type A to universal type O, which we have applied to an Rhnull donor‐derived hiPSC line, originally carrying blood group A. This work provides a paradigmatic example of an approach potentially applicable to other hiPSC lines derived from rare blood donors not carrying blood type O.
The transfusion of red blood cells (RBCs), currently obtained from volunteer blood donations, is an essential therapy for patients with chronic or acute anaemia. This form of cell‐based therapy is an indispensable part of modern healthcare systems. However, the perspective of insufficient blood supply due to population aging, and the potential risk of transfusion‐transmitted infections remain major concerns. , In addition, the scarcity of donors with rare blood types represents a global challenge when compatible red cells with a rare blood phenotype are required for transfusion. , For these reasons, the in vitro generation of RBCs, to supplement the donation system, is nowadays a major focus of research in transfusion medicine. Beyond transfusion requirements, red cells with infrequent phenotypes are also necessary for diagnostic purposes in clinical laboratories. The identification of unusual red cell antibody specificities in patient sera depends on the availability of reagent red cells with rare phenotypes or infrequent antigen combinations for serological crossmatching, which is crucial to allow the accurate selection and effective search of compatible units for transfusion. During the past decade, enormous progress has been made in the in vitro manufacture of human RBCs from different cell sources. , , , , Among these, human‐induced pluripotent stem cells (hiPSCs) provide an unlimited source of hematopoietic progenitor cells, which can subsequently be differentiated into erythroid cells. hiPSC lines can also be derived from easily accessible peripheral blood mononuclear cells (PBMCs) from selected donors , and be amenable to gene editing. , Overall, these features make them a promising source for sustainable production of customized red cells. Different hiPSC lines have already been obtained from existing donors with rare blood types , or have been modified using CRISPR/Cas9 gene editing approaches to reproduce uncommon null phenotypes by knocking‐out specific blood group genes. However, the potential use of hiPSC‐derived red cells is also dependent on the ABO type. Except for the rare Bombay phenotype, extremely infrequent or null blood group types are not necessarily encountered in blood type O donors. Here, we present a CRISPR/Cas9‐mediated ABO gene edition strategy for the conversion of blood type A to universal type O, which we have applied to an Rhnull donor‐derived hiPSC line, originally carrying blood group A. This approach is potentially applicable to other hiPSC lines derived from rare blood donors, not carrying blood type O.
Procedures for sample collection and hiPSC line generation were approved by the Clinical Research Ethics Committee (CEIC) of the Vall d'Hebron Research Institute, the Comisión de Garantías para la Donación y Utilización de Células y Tejidos Humanos (Spanish National Stem Cell Bank, ISCIII) and the Catalan Authority for Stem Cell Research (0336E/10472/2017). Informed consent was obtained from the donor. All procedures involving animals were approved by the Institutional Animal Ethics Board, and the protocols were approved by the Catalan Authority.
To generate Rh null donor‐derived hiPSCs, PBMCs were isolated from a 20 ml‐whole blood sample of the selected blood donor, previously identified and characterized at the Immunohematology Reference Laboratory of the Banc de Sang iTeixits (Barcelona). MNCs were isolated using standard density gradient centrifugation with SepMate™ tubes (StemCell Technologies, Canada). The PBMCs were carefully recovered from the interface and washed in PBS 1×. PBMCs were reprogrammed using the integration‐free CytoTune®‐iPS 2.0 Sendai Reprogramming Kit (ThermoFisher, USA), which contains Sendai virus particles for the expression of the four Yamanaka factors. Undifferentiated hiPSCs were maintained in mTeSR™1 (StemCell Technologies, Canada) on 3 μg/ml of Laminin‐521 (StemCell Technologies)‐coated plates and expanded using EDTA passaging solution (ThermoFisher). Media samples were routinely tested for the absence of mycoplasma contamination using the selective biochemical test MycoAlert™PLUS (Lonza, Switzerland).
For ABO gene targeting, two strategies were approached: (1) the generation of a gene knock‐out (KO) and (2) the insertion of the naturally occurring c.261delG single nucleotide deletion through a short sequence knock‐in (KI). For both strategies, we designed RNA guides (gRNA) using CRISPR‐direct tool (https://portals.broadinstitute.org/gpp/public/analysis‐tools/sgrna‐design). We selected in both cases the target sites with the lowest number of predicted off‐targets. The gRNA sequences are depicted in Table S1. For the KI strategy, we designed a single‐strand donor DNA carrying the c.261delG mutation and a mutated PAM to avoid re‐cutting of the target sequence (Table S1). Each guide was previously transfected with the Alt‐R® S.p. Cas9 Nuclease V3 (IDT# 1081058) and the Alt‐R® CRISPR‐Cas9 tracrRNA, ATTO™ 550, (IDT #1075927) in HEK‐293T cells according to IDT protocol and tested for cutting efficiency by the T7 endonuclease assay (as described below). The most efficient guide was nucleofected with the Cas9 protein into the parental hiPSC line hiPSC#1 using the Neon Electroporation transfection System (ThermoFisher), and cells were plated into geltrex‐coated plates. Nucleofection efficiency was assessed after 24 h. Forty‐eight hours post nucleofection cells were plated as single cells on geltrex‐coated 96‐well plates for clonal selection in mTeSR™1 (StemCell Technologies, Canada) supplemented with 10 μM Y‐27632 and cloneR (StemCell Technologies). Single‐cell clones were expanded for 2 to 3 weeks and subsequently analysed by Sanger sequencing for the target sites modified. To check for unintended targets of our sgRNAs, IDT design checker (https://eu.idtdna.com/site/order/designtool/index/CRISPR_SEQUENCE) was used, and the top five in silico‐predicted off‐targets coding regions of ABO E3 sgRNA (CCDC78, HTR5A, PRRG2, RHBDL2, UCKL1‐AS1) and ABO E6.1 sgRNA (C16orf89, LINC02794, LZTS1, NLCN, SLC8A1) were analysed by sanger sequencing. Sequences are available at https://github.com/anasevilla/ABO‐gene‐editing (see Table S2 for sequencing primers).
HEK‐293T cultures were dissociated with TrypLE express (ThermoFisher) and 2 × 105 cells were transfected by CRISPRMAX Transfection Reagent (ThermoFisher) with 12 pmol of each of the RNAs (gRNA and tracrRNA) and the Cas9 protein. Genomic DNA was extracted 2 days after transfection. Genomic regions flanking the CRISPR target sites were PCR amplified (Table S1). PCR products were denatured, re‐annealed and subsequently treated with 5U of T7EI at 37°C for 15 min.
For RHAG gene sequence analysis, DNA was extracted from donor PBMCs or hiPSCs by an automated method using the QIAsymphony instrument (Qiagen, Germany). Primers to amplify RHAG gene exon 6 are listed in Table S2. For CRISPR edit validation, genomic DNA was isolated from each expanded clone using DNeasy Blood and Tissue Kit (Promega). DNA regions encompassing guide sites were amplified using specific primers (Table S2). Amplification was performed with SequalPrep™ Long PCR Kit with dNTPs (Applied Biosystems, USA). The PCR products were Sanger sequenced using the Big Dye Terminator v1.1 kit (Applied Biosystems). Sequencing primers are also listed in Table S2. DNA sequences were aligned with the reference genomic sequences: NG_011704.1 for the RHAG gene and NG_006669.2 for ABO gene, using the CLC GenomicWorkbench 21.0.3 software (Qiagen).
To confirm the cell line identity, genomic DNA was extracted from iPSC clones as well as donor PBMCs and used for short‐tandem repeat (STR) marker analysis using the Mentype® Chimera®system (Biotype®Diagnostic GmbH, Germany).
Genomic integrity of the generated hiPSCs was evaluated by G‐banded metaphase karyotype analysis (Molecular Citogenetics Laboratory, Hospital del Mar, Barcelona). Briefly, cultures of hiPSCs (70% confluent) were treated with KaryoMaxcolcemid (Invitrogen), dissociated, incubated in hypotonic solution and fixed in Carnoy solution (75% methanol, 25% acetic acid). Karyotyping was performed following standard procedures. A minimum of 15 metaphases were examined.
Total RNA was isolated from hiPSCs and developing embryoid body (EB) cells using the RNeasy Micro kit (Qiagen) and treated with RNase‐free DNase (Qiagen). Total RNA (1 μg) was reverse transcribed using a high‐capacity reverse transcription kit (Applied Biosystems). All quantitative PCR analyses were performed using the Fast SYBR Green Master Mix (Applied Biosystems) following the manufacturer's protocols on the Light Cycler 480 Real‐Time PCR System (Roche). Gene‐specific primers used for this study are listed in Table S3.
hiPSCs were differentiated into hematopoietic progenitor cells (HPCs), using the STEMdiff™Hematopoietic kit (StemCell Technologies) following the manufacturer's recommendations. The 12‐day differentiation protocol was performed in two stages. First, to induce hiPSC commitment towards mesoderm, hiPSC aggregates were plated on laminin‐521‐coated plates and cultured during the first 3 days with STEMdiff™ Hematopoietic Supplement A added to the basal medium. Second, for the subsequent 9 days, mesodermal cells were further differentiated into HPCs using basal medium supplemented with STEMdiff™ Hematopoietic Supplement B performing half‐medium changes at days 5, 7 and 10 according to the manufacturer's instructions. At day 12, HPCs were harvested from the culture supernatant and re‐cultured in erythroid differentiation media. To induce erythroid differentiation, HPCs were cultured in basal medium: IMDM (HyClone™, GE Healthcare Life Sciences, USA) containing 1 U/ml Glutamine, 3% (v/v) human AB serum (Banc de Sang i Teixits, Barcelona, Spain), 2% (v/v) foetal bovine serum (Gibco, USA), 10 μg/ml Insulin (Sigma, USA), 3 U/ml heparin (Sigma, USA), 200 μg/ml Iron‐Saturated Transferrin (R&D Systems, USA) and 1U/ml penicillin/streptomycin (Pen/Strep, Gibco) according to the protocol described by Griffiths et al. In the first differentiation stage (days 0–7), basal medium was supplemented with 10 ng/ml SCF, 1 ng/ml IL‐3 and 3 U/ml EPO (PeproTech, USA); in the second differentiation stage (days 7–11), the basal medium was supplemented with 10 ng/ml SCF and 3 U/ml EPO; in the third differentiation stage (days 11–21), basal medium was supplemented with 3 U/ml EPO and additional transferrin to a final concentration of 500 μg/ml. Cells were cultured into stationary plastic tissue culture flasks at a density of 1–3 × 105 cells/ml at the first stage, 3 × 105 – 1 × 106 cells/ml at the second stage and 1–2 × 106 cells/ml at the third stage by full media changes every other day starting on day 3.
Cell surface marker staining was performed by direct immunofluorescence with conjugated monoclonal antibodies listed in Table S4. Briefly, a sample of 1 × 105 cells was incubated with conjugated antibodies in 50 μl of phosphate‐buffered saline (PBS) (Lonza, USA) containing 1% (v/v) bovine serum albumin (BSA) (Grifols, Spain) (PBS‐BSA) for 30 min at room temperature (RT) with continuous mixing in the dark. Cells were washed twice with PBS‐BSA and then analysed on a MACSQuant flow cytometer using MACSQuantify software (Miltenyi Biotech, Germany). For RhAG staining, a sample of 5 × 105 cells was incubated with the monoclonal antibody LA1818 (kindly provided by Prof. Ellen van der Schoot, Sanquin, Netherlands) for 30 min at RT, and then labelled with FITC‐conjugated goat anti‐mouse IgG and IgM (Becton Dickinson) for 30 min at RT in the dark. Results were analysed on a MACSQuant system (Miltenyi Biotech). Cell viability was assessed by 7‐AAD staining.
Cytospins were prepared at the Hematological Cytology Service (Hospital del Mar, Barcelona). Briefly, a sample of 1 × 104 cells was prepared by centrifuging onto glass slides at 500 rpm for 10 min in a Thermo Scientific Cytospin 4 cytocentrifuge. The slides were stained with May‐Grünwald Giemsa stain (Merck) according to the Hematology Cytology Service's protocol. Cytospins were imaged at 400× using an optical microscope.
Expression of A antigen was analysed by direct fluorescence staining with 50 μg/ml AF488‐conjugated helix pomatia agglutinin (HPA) lectin (ThermoFisher) in living cultured erythroid cells. As positive and negative controls, RBCs from individuals of A and O blood group types were fixed with 4% paraformaldehyde previous direct staining. Nuclei were stained with DAPI. Images were taken using Zeiss Axio Observer Z1 – Apotome inverted fluorescent microscope and analysed using the Image J software.
Bio‐Rad DiaClon Rh‐Subgroups+K ID‐Cards with monoclonal typing reagents for C (RH2), c (RH4), E (RH3), e (RH5) were used for serological detection of RhCE antigens. Bio‐Rad DiaClon ABD‐Confirmation for Donors ID‐Cards with monoclonal typing reagents for RhD: ESD‐1 M and 175‐2, were used for serological detection of the RhD antigen. Briefly, cell suspensions prepared from 1–2 × 106 hiPSC‐derived reticulocytes were pelleted and resuspended in 50 μl of ID‐Diluent 2 (Bio‐Rad Laboratories, Switzerland). Cards were centrifuged as per the manufacturer's instructions. A monoclonal anti‐k (Cellano) reagent (Pelikloon IgM monoclonal Lk1) was also used to detect Cellano antigen expression using Bio‐Rad NaCl, Enzyme test and Cold Agglutinins ID‐Cards (Bio‐Rad Laboratories). Fifty microliters of prepared cell suspension were added to a column followed by 25 μl of the anti‐Cellano antibody. Cards were centrifuged as per the manufacturer's instructions.
To obtain an integration‐free Rhnull hiPSC line, PBMCs from an Rhnull female blood donor were reprogrammed. The donor subject had been previously identified as a homozygous carrier of a single‐base mutation (c.836G > A) in the RHAG gene, leading to the rare Rhnull blood type (ISBT RHAG Blood Group Alleles Table:https://www.isbtweb.org/static/5d593bb0‐02e1‐47a2‐9a8fe3e34df68a5e/ISBT030RHAGbloodgroupallelesv6230‐NOV‐2021.pdf. PBMCs were reprogrammed using integration‐free Sendai virus vectors expressing OCT4, SOX2, KLF4 and cMYC under serum‐free and feeder‐free conditions. Two hiPSC lines were generated from this donor and representative clones, named BST PBiPS6‐SV4F‐9 (abbreviated as hiPSC#1) and BST PBiPS6a‐SV4F‐6 (abbreviated as hiPSC#2), were fully characterized. STR analysis confirmed the identity of both lines when compared to the original PBMCs (Figure 1A). Both hiPSC lines robustly proliferated for more than 20 passages, showing a normal diploid female [XX, 46] karyotype, without any detectable numerical or structural chromosomal abnormalities (Figure 1B). The established Rhnull hiPSC lines displayed hallmarks of pluripotency, being positive for alkaline phosphatase staining (Figure 1C) and enhanced endogenous gene expression of common pluripotency markers (Figure 1D). The stemness of hiPSCs was also verified by immunofluorescence of pluripotency markers in hiPSC colonies from passages 8 to 15 (Figure S1A). Definitive proof of a pluripotent phenotype was shown in in vitro‐directed differentiation assays towards the three germinal layers (Figure S1B) and in in vivo teratoma formation assays (Figure S1C,D). Additionally, the presence of the RHAG gene c.836G > A homozygous mutation was also confirmed by gene sequencing (Figure 1E). These results demonstrate that we have successfully obtained an integration‐free and feeder‐free hiPSC line carrying the genotype that leads to the Rhnull phenotype in derived red cells.
The original A blood type of the Rhnull donor was associated with a heterozygous A2/O1 ABO genotype, which was also confirmed in the resultant hiPSC line by ABO gene sequencing (Figure S2). In order to convert the Rhnull hiPSC line from blood type A to the universal type O, we designed two strategies based on CRISPR/Cas9 technology. The first approach was based on the generation of a KI, mimicking the natural (c.261delG) polymorphism, present in the most common inactive ABO*O.01 (O1) allele. This deletion of guanine in exon 6 causes a frameshift (p.Thr88Profs*31) in a sequence that otherwise is identical to the consensus A sequence (Figure 2A). The second approach relied on the generation of a KO targeting the third exon of the ABO gene, also generating a frameshift. In both approaches, three guide gRNA sequences per CRISPR site were designed, transfected and previously tested on HEK293T cells to evaluate their cutting efficiency through endonuclease T7 assay. In the KI strategy, the best‐performing gRNA, which specifically targeted the exon 6 (Figure 2B), was co‐transfected with the donor DNA sequence containing the 261G deletion, which is used by the DNA repair machinery as the new template after the cut. For the KO generation, the Rhnull hiPSC line was transfected by electroporation with the RNP complex whose gRNA targeted the ABO gene exon 3 (Figure 2C). Different clones were isolated and screened by Sanger sequencing to confirm the ABO gene CRISPR edition. For the KO strategy we checked 26 clones and found five with indels producing a truncated protein (19% efficiency). For the KI strategy, we identified five clones carrying the 261G deletion out of 16 clones screened (31% efficiency) (see data at https://github.com/anasevilla/ABO‐gene‐editing). We then selected two KO (KO‐C31 and KO‐C52) and two KI (KI‐C4 and KI‐C5) clones for further characterization (Figure 2B,C). Using the IDT design checker (https://eu.idtdna.com/site/order/designtool/index/CRISPR_SEQUENCE) software, we analysed the top five in silico‐predicted off‐targets of ABO E3 sgRNA (CCDC78, HTR5A, PRRG2, RHBDL2 and UCKL1‐AS1) and ABO E6.1 sgRNA (C16orf89, LINC02794, LZTS1, NLCN and SLC8A1) by sanger sequencing and found them all consistently unaltered in the four selected clones, demonstrating the specificity of our gene editing strategy (sequences are available at https://github.com/anasevilla/ABO‐gene‐editing). Importantly, cell line identity of all four clones was confirmed by STR analysis (Figure S3A) and also showed normal diploid [XX, 46] karyotypes (Figure S3B). Furthermore, gene‐edited hiPSC clones remained pluripotent after CRISPR/Cas9 gene editing, retained hESC‐like morphology and expressed similar (p > 0.05) RNA (Figure S3C) and protein (Figure S3D,E) levels of the pluripotency markers SSEA3, SSEA4, TRA‐1‐60 and TRA‐1‐81 to the parental Rhnull hiPSC line (Figure S3F). In addition, their pluripotent capacity was tested in vitro through directed differentiation into cell lineages representing all three germ layers through the embryoid body assay (Figure S4A,B). Although expected differences across the iPSC lines were observed for the early lineage differentiation markers (SOX17, T, TUJ1) in the differentiating embryoid bodies, no statistical differences were observed for the expression of the pluripotency markers OCT4, SOX2 and NANOG at the pluripotency state between gene‐edited iPSC lines and the parental lines (Figure S4C). Thus, we have established two CRISPR/Cas9‐mediated gene edition strategies to convert blood type A hiPSC lines to type O with no impact on the stemness potential.
The potential of the ABO‐edited hiPSC lines to differentiate towards the erythroid lineage was evaluated in parallel with the parental Rhnull hiPSCs in three independent experiments. The parental Rhnull hiPSCs and the KI‐C5 and KO‐C52‐edited clones were first differentiated towards HPCs with the STEMDiff™ Hematopoietic Kit (Figure S5A). At day 12 of the differentiation protocol, HPCs released from hematopoietic clusters were harvested from the culture supernatant (Figure S5B). This population contained around 90% CD34+ cells, and around 60% of these cells were CD45low/+ (Figure S5C). To further characterize the CD34+CD45low/+ fraction, we also analysed the expression of the erythroid lineage surface markers CD71, CD235a, CD49d and CD233. We confirmed the presence of a variable range (36–75%) of early erythroblasts CD71+CD235a+CD49d+CD233− (Figure S5C). The collected cells, containing erythroid‐committed HPCs, were further cultured in erythroid differentiation medium according to the three‐step protocol described in Materials and Methods (Figure 3A). The follow‐up of cell viability and expansion throughout the culture showed no statistically significant differences between the edited and the parental cell lines regarding the survival/proliferative capacity of hiPSC‐derived erythroid progenitors (Figure S6A,B). Distinct stages of erythroid maturation were assessed morphologically at four time points (d0, d7, d14 and d21) (Figure 3B). The erythroid cells progressed through distinct erythroid stages with orthochromatic erythroblasts already appearing at day 7, showing no statistically significant differences between the parental Rhnull hiPSC and the edited clones, KI‐C5 and KO‐C52, across the 21‐day differentiation period (Figure S6C). At day 21, orthochromatic erythroblasts were the predominant cells (approximately 80%), with a very low proportion of enucleated cells (6–8%), in both the parental Rhnull hiPSC line and the edited clones (Figure 3C). Moreover, erythroid differentiation was also assessed by flow cytometry immunophenotyping of erythroid surface markers: CD44, CD49d (α4‐integrin), CD71, CD235a (glycophorin A, GPA), CD233 (Band3) and CD238 (KEL). The observed dynamic changes of expression revealed an analogous progression through erythroid differentiation in parental Rhnull hiPSC line and the edited clones. In brief, we observed two distinct patterns of expression (Figure 3D). The adhesion molecules CD44 and CD49d, as well as the transferrin receptor CD71 presented a pattern with high levels of expression in early‐stage erythroblasts and a progressive decrease in late‐stage erythroblasts. In contrast, a different pattern was observed for the CD235a, CD233 and CD238 markers, which displayed low levels of expression in early‐erythroblasts with a progressive increase in late‐stage erythroblasts. Our results concur with the expected progression of erythroid cell differentiation cultures from hiPSC‐derived CD34+ HPCs, , , with no significant differences between the parental Rhnull hiPSCs and the edited clones (Figure S6D).
To confirm the Rhnull phenotype, we first assessed RhAG expression by flow cytometry, in cells differentiated from both the parental and ABO‐edited hiPSCs, using the LA1818 anti‐RhAG monoclonal antibody. No RhAG expression was observed on the membrane of erythroid cells derived either from the parental Rhnull hiPSCs or from the ABO‐edited clones KI‐C5 and KO‐C52 (Figure 4A). As the RhAG glycoprotein is essential for the Rh complex formation, we next assessed the expression of the Rh antigens (D, C, c, E and e) by agglutination tests using gel card technology, and no agglutination was observed with any of the anti‐Rh typing reagents (Figure 4B), further confirming the Rhnull phenotype. These data confirm the successful generation of in vitro differentiated erythroid cells reproducing the Rhnull phenotype of the original donor subject, which has neither been affected by the reprogramming of the donor's PBMCs, nor by the CRISPR edition of the ABO gene.
To analyse A antigen expression in differentiated erythroid cells from parental Rhnull hiPSCs and edited clones, fluorescence labelling was performed using HPA, a lectin that has anti‐A human blood group specificity. The erythroid cells derived from ABO‐edited clones were negative for A antigen expression in flow cytometry studies (Figure 4C). Similarly, no differences in H antigen expression were detected between the parental Rhnull hiPSCs, expressing blood type A2, and the edited lines (blood type O) (Figure 4D). This result indicates that H antigen, which is the precursor for A and B antigen synthesis, is likewise expressed in erythroid cells differentiated from both the Rhnull and edited hiPSCs. In the parental Rhnull hiPSC line, we observed early A antigen expression at day 7 of erythroid differentiation and its maintenance until day 21 (Figure 4E). As expected, though, cultured red cells showed weak A antigen expression, in agreement with the original A2 subgroup of the Rhnull donor (Figure S7). In contrast, we could not detect A antigen‐labelled cells in those cultures differentiated from any of the edited clones. These results confirm the successful conversion of blood type A to blood type O using CRISPR‐Cas9 editing strategies in hiPSCs carrying the rare Rhnull blood group.
Red cells with rare blood types are currently in limited supply due to the scarce representation of these phenotypes in the global population. The provision of compatible red cells for the transfusion of immunized patients carrying rare blood types is one of the first potential target applications of human red cells manufactured in vitro. On the other hand, the diagnosis of red cell alloimmunization, crucial to ensure the safe transfusion of immunized patients, relies on using carefully selected reagent red cells with well characterized phenotypes. These red cells are also obtained from blood donors, so the availability of infrequent phenotypes needed to properly identify rare antibody specificities is likewise very limited. In this sense, having an alternative (unlimited) cell source to derive red cells with rare blood types in vitro, could potentially overcome the current rare blood limitations in both transfusion and diagnostics. One of the approaches that have been considered to address this issue, is the use of hiPSCs obtained from existing donors or patients with rare blood types. , Of course, such donors are not easily available in practice, as rare phenotypes are usually found in less than 1 per 1,000 in the general population, and donors with exceptional ‘null’ phenotypes are even much less represented. One such example is the H deficiency, also known as the Bombay (Oh) phenotype, which is found in 1 in 10,000 individuals in India. The generation of hiPSCs from the dermal fibroblasts of a Bombay blood‐type individual provided the first proof of concept for this approach. More recently, hiPSC lines have been obtained by reprogramming erythroid progenitors from peripheral blood of individuals with the Jr(a−) and D− rare blood types, demonstrating the feasibility of producing autologous hiPSC‐derived red cells for the transfusion of patients with rare blood groups. However, the potential utility of hiPSC‐derived red cells with null phenotypes or infrequent antigen combinations, extends beyond an autologous use. Such hiPSC lines could provide cultured red cells with difficult‐to‐supply blood types for the transfusion of certain groups of immunized patients (e.g., sickle cell disease patients). Likewise, hiPSC lines could solve the limited availability of reagent‐red cells with rare phenotypes, which are also necessary for the identification of rare RBC antibody specificities. In this study, we present the generation of a hiPSC line derived from an Rhnull donor with blood type A. The donor subject had been previously identified as a homozygous carrier of the c.836G > A single‐base mutation in the RHAG gene, leading to the rare Rhnull blood type. , , The Rh‐blood group deficiency, or Rhnull, is an extremely rare phenotype which lacks all the Rh antigens on the red cell membrane. Such valuable blood is necessary for the transfusion support of Rh immunized patients, not only those with Rhnull phenotype but also patients with antibodies against any high‐prevalence Rh specificity, for whom compatible blood is always difficult to procure. Nonetheless, the potential use of hiPSC‐derived red cells for both, transfusion and diagnostics, is also dependent on ABO compatibility. Extremely infrequent or null blood group types are not necessarily encountered in blood type O donors, as it is the case in this blood type A Rhnull donor. This circumstance limits the potential use of the hiPSC‐derived red cells due to the naturally occurring ABO hemagglutinins. To overcome this limitation, we considered the conversion of blood type A to universal type O. The conversion of blood group types A and B to universal type O has been pursued for a long time through approaches based on enzymatic treatment. , , Blood types A and B differ from type O in the presence of an additional sugar residue (GalNAc or Gal, respectively) on the precursor H‐antigen found on type O RBCs. The concept of removing these immunogenic sugars by specific enzymes (glycosidases) from blood type A or B red cells, was first proposed and demonstrated by Goldstein (1982). The first attempts required massive amounts of enzyme but novel α‐galactosidases and α‐N‐acetylgalactosaminidases have been shown to improve the conversion efficiency. However, this technology has not yet moved into clinical practice, as there are hold‐ups pending to be solved. Alternatively, we addressed the conversion of the blood type A Rhnull hiPSC line into universal type O using CRISPR/Cas9‐mediated gene editing technology, which allows the precise, robust and efficient edition of genes of interest. With the aim to abrogate the expression of the α‐1,3‐N‐acetylgalactosamine transferase (A‐transferase), we designed two different ABO gene edition approaches. The first approach was based on the generation of a KI, mimicking the c.261delG single nucleotide deletion, present in the most common inactive ABO*O.01 (O1) allele. The specific and precise incorporation of this c.261delG polymorphism within the ABO gene has been attempted in the present work aiming to reproduce the genetic basis naturally associated to blood type O. Exploiting the CRISPR/Cas9‐targeted integration to correct genetic defects has led to a number of proof‐of‐principle works in patient‐derived hiPSCs, in which the mutations responsible for cystic fibrosis, haemophilia A and β‐thalassemia were successfully corrected, although with a limited efficiency. , , Thus, a second approach by gene KO was undertaken in parallel to maximize the possibilities to achieve our final goal, which was to obtain an Rhnull hiPSC line converted to universal blood type O. Indeed, both the KI and the KO strategies successfully rendered hiPSC‐edited clones with the intended ABO gene modifications, as demonstrated by ABO sequencing analysis. Moreover, the established ABO‐edited hiPSCs lines maintained the Rhnull‐related RHAG gene mutation as well. The results obtained in the characterization of the KI and KO‐edited hiPSC lines showed no changes on their stemness potential. Likewise, these lines have been successfully differentiated into HPCs and, subsequently, to the erythroid lineage. No remarkable differences have been observed between the parental Rhnull hiPSCs line and the edited clones in erythroid differentiation experiments, with overall results concording with the expected progression of erythroid cell differentiation cultures from hiPSC‐derived CD34+ progenitor cells. , , It is worth noting that we have been able to produce differentiated erythroid cells reproducing the Rhnull phenotype, proving no alteration of the original donor's rare phenotype due to the PBMCs reprograming or to the subsequent hiPSCs CRISPR ABO‐gene edition. Remarkably, the results obtained from both KI and KO gene edition strategies provide the first demonstration of blood type A conversion to the universal type O using CRISPR/Cas9 technology. The knock‐out of specific blood group genes, other than ABO, in pre‐existing hiPSC lines has been recently reported as an strategy to reproduce uncommon null phenotypes. Here, we demonstrate the feasibility of robust and sustainable ABO blood type conversion using these newly designed CRISPR/Cas9 gene editing approaches, allowing the production of cultured red cells with improved ABO compatibility. The potential application of these approaches is not restricted to hiPSC lines, since they can also be applied to other cell lines of interest for cultured red cells production, such as immortalized human erythroblast cell lines, , derived from individuals not carrying blood type O. During the past decade, significant advances have been made in the production of manufactured red cells from different cell sources. , , , Despite the known limitations that still need to be overcome (e.g., low enucleation rate and cost‐efficient scaling), the deeper knowledge of the regulatory pathways involved in terminal erythroid differentiation, together with the continuous progress in scaled‐up protocols and technological achievements, allow to anticipate that in vitro production of RBCs will be possible in the near future. In this context, CRISPR/Cas9‐mediated blood group gene edition will certainly play an important role as a tool to improve blood group compatibility, like this work demonstrates.
The authors declare no competing interests.
The data that support the findings of this study are openly available at https://github.com/anasevilla/ABO‐gene‐editing and in the Supplementary Files.
Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. Click here for additional data file. | true | true | true |
|
PMC9594343 | 36162812 | Mario Escobar,Jing Li,Aditi Patel,Shizhe Liu,Qi Xu,Isaac B. Hilton | Quantification of Genome Editing and Transcriptional Control Capabilities Reveals Hierarchies among Diverse CRISPR/Cas Systems in Human Cells | 26-09-2022 | genome editing,CRISPR/Cas systems,gene regulation,CRISPRa,CRISPRi | CRISPR/Cas technologies have revolutionized the ability to redesign genomic information and tailor endogenous gene expression. Nevertheless, the discovery and development of new CRISPR/Cas systems has resulted in a lack of clarity surrounding the relative efficacies among these technologies in human cells. This deficit makes the optimal selection of CRISPR/Cas technologies in human cells unnecessarily challenging, which in turn hampers their adoption, and thus ultimately limits their utility. Here, we designed a series of endogenous testbed systems to methodically quantify and compare the genome editing, CRISPRi, and CRISPRa capabilities among 10 different natural and engineered Cas protein variants spanning Type II and Type V CRISPR/Cas families. We show that although all Cas protein variants are capable of genome editing and transcriptional control in human cells, hierarchies exist, particularly for genome editing and CRISPRa applications, wherein Cas9 ≥ Cas12a > Cas12e/Cas12j. Our findings also highlight the utility of our modular testbed platforms to rapidly and systematically quantify the functionality of practically any natural or engineered genomic-targeting Cas protein in human cells. | Quantification of Genome Editing and Transcriptional Control Capabilities Reveals Hierarchies among Diverse CRISPR/Cas Systems in Human Cells
CRISPR/Cas technologies have revolutionized the ability to redesign genomic information and tailor endogenous gene expression. Nevertheless, the discovery and development of new CRISPR/Cas systems has resulted in a lack of clarity surrounding the relative efficacies among these technologies in human cells. This deficit makes the optimal selection of CRISPR/Cas technologies in human cells unnecessarily challenging, which in turn hampers their adoption, and thus ultimately limits their utility. Here, we designed a series of endogenous testbed systems to methodically quantify and compare the genome editing, CRISPRi, and CRISPRa capabilities among 10 different natural and engineered Cas protein variants spanning Type II and Type V CRISPR/Cas families. We show that although all Cas protein variants are capable of genome editing and transcriptional control in human cells, hierarchies exist, particularly for genome editing and CRISPRa applications, wherein Cas9 ≥ Cas12a > Cas12e/Cas12j. Our findings also highlight the utility of our modular testbed platforms to rapidly and systematically quantify the functionality of practically any natural or engineered genomic-targeting Cas protein in human cells.
Significant phylogenetic and functional diversity has recently been uncovered among CRISPR/Cas systems. Remarkably, components from at least 6 different CRISPR/Cas families (Cas9, Cas12a, Cas12b, Cas12e, Cas12f, and Cas12j) have been identified that enable editing of mammalian genomes, transcriptomes, and/or epigenomes (Figure S1). Although the Type II CRISPR/Cas system identified in Streptococcus pyogenes (SpCas9) is the most well-characterized and widely used system for genome editing, CRISPR activation (CRISPRa), and CRISPR interference (CRISPRi) in mammalian cells, the relatively large size of SpCas9, the NGG PAM sequence requirements, and the potential for off-target effects can restrict the utility of SpCas9 in some contexts. To overcome these limitations, engineered variants of SpCas9 with increased fidelity, such as hypoCas9, eSpCas9, and HiFi Cas9, have been developed for use in mammalian cells. Additionally, SpCas9 variants with altered PAM specificities, such as SpRYCas9, VQR SpCas9 and EQR SpCas9, SpCas9-NG, and xCas9, have been created. Furthermore, natural Cas9 orthologues, such as Staphylococcus aureus Cas9 (SaCas9);Neisseria meningitidis Cas9 (NmCas9); and Campylobacter jejuni Cas9 (CjCas9), have been characterized that have smaller sizes and different PAM requirements than SpCas9, yet still retain robust activity in mammalian cells. CRISPR/Cas systems from the Type V family, such as Cas12a (also called Cpf1), Cas12e (also called CasX), and Cas12j (also called CasΦ), have also been adopted for use in mammalian cells. Unlike Type II CRISPR/Cas family members, Cas proteins associated with Type V CRISPR/Cas systems contain a single nuclease domain, require a T-rich PAM, and typically have smaller sizes than SpCas9. Although multiple CRISPR/Cas systems have been identified and tested in mammalian cells, differing PAM sequence requirements and the lack of isogenic expression vectors has made direct, systematic comparisons among genome editing and transcriptional modulatory (i.e., CRISPRa/CRISPRi) activities in mammalian cells challenging. Additionally, for many of the newly identified CRISPR/Cas nucleases, no corresponding CRISPRa or CRISPRi tools have been developed. Such tools could be particularly useful for endogenous gene activation or repression strategies in combination with smaller CRISPR/Cas systems (i.e., Cas12e and Cas12j). Here, we selected 10 different CRISPR/Cas variants from both Type II and Type V families and developed a series of genome editing and CRISPRa/CRISPRi tools in isogenic expression vector backbones. We quantified the genome editing efficacy of each variant using different gRNA-targeted sites within an integrated eGFP testbed in the human HEK293T cell line. We also measured the genome editing efficacy of these different CRISPR nucleases at the endogenous EMX1 locus in HEK293T, HeLa, and U2OS cells. Additionally, we evaluated the CRISPRi (using 10 different KRAB fusion proteins) and CRISPRa (using 10 different VPR fusion proteins) capabilities of each variant at endogenous human loci. Further, we designed integrated testbed frameworks to benchmark any of the selected CRISPRi/CRISPRa orthologues or variants at the same target site regardless of the respective PAM requirements. We find that across these human cell lines, Type II CRISPR/Cas systems generally outperform Type V systems in both genome editing and transcriptional activation, and that nearly all CRISPR/Cas systems permit transcriptional repression when fused to the KRAB domain. Collectively, our studies clarify the relative efficacies of diverse natural and engineered CRISPR/Cas systems in human cells and provide a useful set of new ready-to-use expression vectors and assay testbeds that can enable rapid and robust in situ comparisons among current and future genomic-targeted CRISPR/Cas variants.
The CRISPR/Cas-based genome editing toolbox has rapidly expanded in recent years, which in turn has established that Cas protein variants with different sizes (ranging in size from ∼400 to ∼1300 amino acids) and different PAM sequence specificities can function in human cells. Despite this exciting progress, systematic analyses to quantify the relative genome editing efficacies among these variants in human cells are lacking, particularly for newly described variants from the Cas12e and Cas12j families. To evaluate and compare the relative genome editing activities of these Cas protein variants, we selected 10 Cas enzymes (Figure 1A) including natural and engineered variants from the Cas9 and Cas12a families (SpCas9, HiFi Cas9, SpRY Cas9, SaCas9, AsCas12a, and LbCas12a), and variants from the more recently described Cas12e and Cas12j families (DpbCas12e, PlmCas12e, Cas12j2, and Cas12j3). We cloned each of these Cas proteins into an isogenic expression vector backbone such that each Cas protein harbored a nuclear localization sequence (NLS), a FLAG epitope tag, and was transcribed by the core EF1α shortened (EFS) promoter (Figure 1B). We transiently co-transfected each Cas variant-encoding vector along with a second vector expressing a corresponding, species-matched gRNA scaffold and a non-targeting control protospacer (Table S1) into HEK293T cells to first evaluate relative expression levels among Cas protein variants in human cells using Western blotting. Interestingly, despite controlled transfection conditions, isogenic vector designs, and reported optimizations for all tested Cas variants in human cells, differing expression levels were observed 72 h post-transient co-transfection in HEK293T cells (Figures 1C and S2). For instance, although Cas12a and Cas9 family members were generally well expressed, Cas12e and Cas12j family members were relatively poorly expressed. In fact, Cas12e and Cas12j variants were only detectable via Western blotting with higher total protein loading amounts (∼50 μg) and longer exposure times (∼600 s; Figure 1D). Flow cytometry to detect the FLAG epitope on the C-terminus of each nuclease active Cas variant in transfected HEK293T cells recapitulated our Western blotting results (Figure S2). Interestingly, the differences between the expression levels of these Cas variants in HEK293T cells were not due to differences in the amounts of respective plasmids transfected into cells (Figure S3).
We next tested the relative efficacies among these 10 selected Cas variants for targeted insertions and deletions (indels) in HEK293T cells using an eGFP disruption assay (Figure 2A). Multiple gRNAs targeting eGFP were designed for each Cas variant to maximize comparative analysis. gRNAs were also selected to target as closely as possible to one another (given PAM sequence restrictions). For Cas12a variants, only two different gRNAs targeting eGFP were available to test disruption efficacy, whereas for all other variants we tested disruption efficacy using three different gRNAs targeting eGFP (Figures 2B and S4). eGFP was integrated into a master HEK293T cell line at a multiplicity of infection (MOI) of ∼10.0 and then eGFP positive cells were sorted, collected, and tested as bulk cell populations. Bulk cells were used to ensure that the eGFP expression was equivalently distributed across all experimental conditions. Similarly, an MOI of ∼10.0 was used to normalize chromatin landscapes at integration sites across all experiments. We also introduced a PEST domain on the C-terminus of eGFP to reduce eGFP background/half-life, similar to previously described assay designs. Despite variable expression levels among tested Cas variants, all Cas enzymes were able to significantly disrupt the eGFP expression at all gRNA-targeted sites relative to non-targeting gRNA control-treated HEK293T cells (Figure 2C). Consistent with previous reports, SaCas9 displayed comparable, yet measurably higher genome editing efficacy compared to SpCas9. Regardless, SpCas9 and SaCas9 both displayed slightly better nuclease activities than Cas12a variants (AsCas12a and LbCas12a). Further, Cas12j family enzymes (Cas12j2 and Cas12j3) exhibited the lowest nuclease activities in this controlled testbed system (Figure 2C). Finally, as has been observed previously at some loci, the engineered HiFiCas9 and SpRYCas9 variants (derived from WT SpCas9), displayed slightly lower indel rates than WT SpCas9. To extend our analysis beyond this eGFP testbed system, we targeted each Cas nuclease to the endogenous EMX1 locus in HEK293T, HeLa, and U2OS cells (Figure S5A). Tracking of indels by decomposition (TIDE) analysis at EMX1 largely replicated the trends observed at the testbed locus (Figure S5B). These trends were most consistent among HEK293T and HeLa cells, and to a lesser extent in U2OS cells. Regardless, these data indicate that although there may be a hierarchy of nuclease efficacy among Cas variants (i.e., Cas9 ≥ Cas12 > Cas12j) in human cells, each enzyme is capable of effective genome editing within human cells.
The intrinsic nuclease activity of Cas proteins can be deactivated through mutagenesis of catalytic amino acid residues. These nuclease-deactivated Cas (dCas) proteins have revolutionized the ability to alter the endogenous human epigenome and/or activate or repress human genes using CRISPRa and CRISPRi approaches, respectively. Although targeting a dCas protein to a human promoter can, in some cases, result in reduced downstream gene expression, this inhibitory effect is more consistent, and often amplified, by fusing a Krüppel-associated box (KRAB) domain to the C-terminus of the dCas9 protein. Therefore, we fused a KRAB domain to the C-terminus of 10 nuclease-deactivated Cas9, Cas12a, Cas12e, or Cas12j proteins in isogenic vector backbones (Figure 3A) to quantify the relative abilities of variants from these families to repress gene expression in human cells. Interestingly, we observed that the fusion of a KRAB domain to the C-terminus of Cas proteins resulted in slight changes to their relative expression in HEK293T cells (Figure S6). Nevertheless, when targeted to the endogenous human CXCR4 promoter in HEK293T or U2OS cells (Figure 3B), all dCas-KRAB variants displayed significant (P < 0.05) repression of CXCR4 transcription relative to mock transfected control cells (Figure 3C, top and bottom). In HeLa cells, all dCas-KRAB variants were also capable of significant (P < 0.05) repression of CXCR4 transcription relative to mock transfected control cells with the exception of dCas12j2-KRAB (Figure 3C, middle). This trend was largely consistent across gRNAs targeting either forward or reverse genomic strands, although effects related to gRNA orientation and/or relative distance from the CXCR4 TSS were evident for some Cas variants. Although we designed CXCR4-targeting gRNAs for each respective dCas-KRAB variant to be as close as possible to one another, an exact target overlap within the native human genome is currently impossible due to differing PAM sequence requirements. Therefore, to more precisely compare the repressive capacities of each dCas-KRAB variant, we created four equivalent CRISPRi lentiviral testbeds that varied only in terms of which PAM sequence (5′-GGGAGT-3′ for Cas9 proteins, 5′-TTTC-3′ for Cas12a proteins, 5′-TTCA-3′ for Cas12e, and 5′-TTA-3′ for Cas12j proteins, respectively) was placed upstream of an EFS promoter constitutively driving eGFP (Figures 4B and S7A). All PAMs were next to a synthetically introduced protospacer without predicted genomic off-targets, which enabled us to target each CRISPRi tool to the same exact sequence/spacing upstream of the integrated EFS promoter. Each respective eGFP expressing testbed was integrated into HEK293T cells at an MOI of ∼10.0 to generate 4 master HEK293T cell lines (Figure 4A) and further, eGFP positive cells were sorted, collected, and tested as bulk cell populations. As above, these steps were taken to equilibrate distribution of eGFP expression and normalized chromatin landscapes at integration sites across all experiments. Similar to our results at the endogenous CXCR4 locus, all tested dCas-KRAB variants significantly (P < 0.05) repressed eGFP expression when targeted 16bp upstream of the EFS promoter in testbed systems in HEK293T cells (Figures 4C and S8). These data demonstrate that all tested nuclease-inactivated CRISPR/Cas systems from Type II and Type V families are capable of human gene repression when fused to the KRAB domain and that our testbed CRISPRi platform enables rapid, simple, and robust quantification of the relative efficacy of diverse CRISPRi tools or even repressive dCas-based epigenome editing technologies in future iterations. dCas proteins have also been used to activate endogenous genes in so-called CRISPRa settings wherein the dCas protein is used as a scaffold to recruit transcriptional activation domains to regulatory elements such as promoters or enhancers. One such CRISPRa system leverages a tripartite transcriptional activation domain called VPR (VP64-p65-Rta) fused the C-terminus of a dCas protein. To measure the relative abilities of dCas variants to activate transcription in human cells, we fused the VPR domain to the C-termini of the selected dCas9, dCas12a, dCas12e, or dCas12j proteins in the isogenic vector backbones (Figure 5A). As observed with the KRAB domain, fusion of the VPR domain to the C-terminus of different dCas proteins resulted in reduced protein expression in HEK293T cells (Figure S9). Despite this effect, when targeted to the human IL1RN promoter (Figure 5B) in HEK293T and HeLa cells, all dCas-VPR variants displayed the ability to activate endogenous IL1RN expression, albeit relatively inconsistently among variants (Figure 5C, top and middle). In U2OS cells, although Type II dCas-VPR variants were capable of significant (P < 0.05) activation of IL1RN transcription relative to mock transfected control cells, Type V dCas-VPR fusions were markedly less effective (Figure 5C, bottom). Together this indicates that overall dCas9/dCas12a-VPR fusions were more effective at activating IL1RN gene expression than dCas12e/dCas12j-VPR fusions in these three human cell lines. Again, because targeting each respective dCas-VPR variant to the same exact sites within the native human genome is currently impossible due to differing PAM sequence requirements, we created four equivalent CRISPRa lentiviral testbeds (Figure S7B) to more systematically compare the transactivation capacities of selected dCas-VPR variants (Figure 6A). Similar to our CRISPRi testbeds, each CRISPRa lentiviral testbed varied only in terms of which PAM sequence (5′-GGGAGT-3′ for Cas9 proteins, 5′-TTTC-3′ for Cas12a proteins, 5′-TTCA-3′ for Cas12e, and 5′-TTA-3′ for Cas12j proteins, respectively) was placed upstream of a target promoter that could drive the eGFP expression when stimulated (Figure 6B). However, for these CRISPRa testbed experiments, the miniCMV promoter was selected to drive eGFP in response to dCas-VPR variants because it has been found to display low basal expression and to be highly responsive to the VPR domain in human cells. All PAMs were designed next to a synthetically introduced protospacer without predicted genomic off-targets, which enabled us to target each CRISPRa tool to the same exact sequence/spacing upstream of the integrated miniCMV promoter. As shown above, each respective miniCMV-eGFP testbed was integrated into HEK293T cells at an MOI of ∼10.0 to generate 4 master HEK293T cell lines (Figure 6A). Transduced cells were tested as bulk cell populations to ensure an equivalent distribution of eGFP expression and normalized chromatin landscapes at integration sites across all experiments. All dCas-VPR variants significantly (P < 0.05) upregulated eGFP expression when targeted 16bp upstream of the miniCMV promoter (Figure 6C). However, overall Cas9 family enzymes performed the best in CRISPRa testbed assays (up to ∼30-fold activation) followed by Cas12a family variants (up to ∼15 fold activation). Although the CRISPRa tools based on Cas12e and Cas12j family variants only slightly activated eGFP expression (up to ∼3 fold), this was nonetheless significantly (P < 0.05) above non-targeting gRNA control-treated HEK293T cells. Collectively, these results suggest that all nuclease-inactivated CRISPR systems are capable of human gene activation when fused to the VPR domain but that Cas12e and Cas12j family variants are less effective than Cas9 and Cas12a systems. Finally, these data support the use of our testbed CRISPRa platform as a robust and straightforward framework to rapidly evaluate and benchmark CRISPRa technologies. The recent expansion of CRISPR/Cas-based tools available for use in human cells has transformed the ability to reshape the human genome, transcriptome, and epigenome. Nevertheless, a lack of clarity exists surrounding the relative endogenous efficacies of these powerful technologies. Consequently, the selection of an optimal CRISPR/Cas system for a particular application has been unnecessarily challenging, and this challenge in turn has limited the adoption of otherwise extremely useful synthetic biology technologies. Here, we designed a series of robust integrated testbeds to systematically compare the genome editing, CRISPRi, and CRISPRa capabilities of both the most commonly used, and newly developed, CRISPR/Cas systems to address this lack of clarity. Our results demonstrate that although all Cas protein variants are generally capable of genome editing and transcriptional control, some variants outperform others in human HEK293T cells. We also built 20 different Cas fusion proteins in isogenic expression vector backbones that can be used for CRISPRi and CRISPRa applications across different PAM targeting and specificity landscapes. Our findings here highlight the utility of our optimized testbed platforms to limit the variables that can confound effective comparisons among current, and future CRISPR/Cas variants in human, or other cell types. Our results here using these testbeds reveal an apparent hierarchy among Cas variants in terms of genome editing and CRISPRa efficacies in human cells. Specifically, Cas9 variants (i.e., WT SpCas9, SpCas9 derivatives, and WT SaCas9) generally perform comparably to one another and perform as well as, or better than, Cas12a variants. However, Cas9 and Cas12a systems both outperform Cas12e or Cas12j variants in our experimental systems (Figures 2, 5, and 6). Interestingly, although Cas12e and Cas12j variants showed moderate genome editing activity (Figures 2C and S5B), each variant displayed only weak gene activation capabilities at both endogenous and synthetic loci when fused to the VPR effector domain (Figures 5C and 6C). This result is consistent with a recent study focusing on another small Type V CRISPR/Cas family member; Cas12f (also called Cas14), which found that dCas12f-VPR could only activate human genes after substantial engineering of the corresponding gRNA and dCas12f protein (the CasMini system). These results suggest that there is a high likelihood that protein and/or gRNA engineering efforts could also improve the genome editing and/or CRISPRa activities of Cas12e and Cas12j variants in human cells, although by design our goal here was to benchmark selected Cas variants with minimal manipulation relative to published compositions. In contrast to the variation observed in genome editing and CRISPRa activities, all tested CRISPR/Cas systems displayed relatively consistent CRISPRi potencies at both endogenous and testbed targets when fused to the KRAB domain (Figures 3C and 4C). This difference between CRISPRi (using the KRAB domain) and CRISPRa (using the VPR domain) effects could be due to the relatively small size of the fused KRAB domain (∼65aa) compared to the VPR domain (∼531aa). Alternatively (and not mutually exclusive), these variations could be due to intrinsic mechanistic differences between endogenous gene activation vs gene repression. Interestingly, our data demonstrate that although the fusion of transcriptional effectors to Cas proteins can reduce their relative expression, and the magnitude of this reduction can be more dramatic for VPR fusions than for fusions containing the KRAB effector (Figures S6 and S9), the tools nonetheless generally retain respective functionality. Whether engineering efforts that result in increased expression would translate into increased CRISPRi/a, or even CRISPR nuclease, potencies, remains to be determined. Notably, the inconsistencies observed among Cas variants could also be driven by differences in relative gRNA/crRNA binding affinities among different Cas proteins. For instance, the crRNA binding affinity of Cas12a has been observed to be lower than that of Cas9, which could partly explain why Cas12a (or dCas12a-VPR fusions) might not edit (or activate) genes as robustly as Cas9 family enzymes (or dCas9-VPR fusions). Additionally, gRNA/crRNA activity could be gRNA/crRNA sequence dependent. Regardless, our results show that Cas12a is an effective endonuclease in human cells, and given the smaller size, potential for multiplexing, reports of reduced off-targeting, and findings that the Cas12a system is amenable to engineered enhancement, it, therefore, is a very promising emergent technology. Similarly, the KRAB/VPR fusions based on SadCas9 generally showed a comparable efficacy in CRISPRi/a relative to SpdCas9-based tools (Figures 3C, 4C, 5C, and 6C). Moreover, SaCas9 performed slightly better than SpCas9 in our disruption assays (Figures 2C and S5B). Therefore, SaCas9 derivatives are excellent alternatives to SpCas9 proteins, especially in applications where the payload size is restricted, such as AAV delivery. Altogether, our studies demonstrate that CRISPR/Cas systems from diverse families are useful for applications in genome editing and transcription modulation within human cells. However, using our controlled isogenic expression cassettes, we found that different Cas variants can have inconsistent expression levels, at least in HEK293T cells. Furthermore, we find that the genome editing and CRISPRa efficacies are not equivalent among different CRISPR systems at both endogenous sites and at carefully designed testbed systems wherein experimental variables, aside from which Cas variant is tested, were minimized. In fact, our data using these tightly controlled experimental settings indicate that there is a hierarchy for relative genome editing and CRISPRa activities among the 10 different Cas variants that we tested here in which Cas9 performs better than or as good as Cas12a and both Cas9 and Cas12a outperform Cas12e/Cas12j. Given the biological nuances distinguishing different Cas variants, there is simply no system with which to benchmark them perfectly. However, our testbed frameworks described here provide a powerful and effective method to evaluate the relative activities of current, or future CRISPR/Cas systems that target genomic DNA in practically any mammalian cell that can be transduced or transfected. Therefore, as more CRISPR/Cas systems continue to be discovered and optimized for use in mammalian/human cells, the technologies that we have described here can be used to rapidly and quantitatively compare their endogenous efficacies. These testbeds could also be useful in screening for improved functions of engineered Cas proteins in endogenous contexts in a high throughput. Collectively, our studies and these new quantitative capabilities can help foster the adoption, implementation, and improvement of the rapidly expanding CRISPR/Cas-based toolbox and thereby augment their utility in providing innovative opportunities across basic and applied biomedical research.
18 Cas protein sequences were retrieved from Genbank and/or previous reports then analyzed using the multiple sequence alignment program MUSCLE. Phylogenetic tree topology diagrams were generated using the Maximum Likelihood method via the MEGA X software package. Sequence identities for SpCas9, SaCas9, NmCas9, AsCas12a, LbCas12a, FnCas12a, DpbCas12e, PlmCas12e, Cas12j2 (CasΦ2), and Cas12j3 (CasΦ3) were generated from WP_032462936.1, AXB99496.1, MBH2503069.1, WP_021736722.1, QRU95066.1, WP_216372291.1, OGP07438.1, OHB99618.1, PDB: 7LYS_A, and PDB: 7ODF_A, respectively. Sequence identity for Cas12j1 (CasΦ1) was derived from a previous report.
All plasmids encoding Cas protein variants and all testbed vectors constructed in this work are available through Addgene. SpCas9, HiFi Cas9, SpRY, AsCas12a, LbCas12a, DpCas12e, PxCas12e, Cas12j2, and Cas12j3 nucleases were polymerase chain reaction (PCR)-amplified from lentiCRISPR v2 (Addgene, 52961), pX165-HiFi Cas9 (Addgene, 140563), pCMV-T7-SpRY-P2A-EGFP (RTW4830; Addgene, 139989), pY026 (Addgene, 84741), pY027 (Addgene, 84742), pBLO 62.4 (Addgene, 123123), pBLO 62.5 (Addgene, 123124), pPP441 (Addgene, 158801), and pPP444 (Addgene, 158802), respectively. SaCas9 sequence was based on pX600-AAV-CMV:NLS-SaCas9-NLS-3xHA-bGHpA (Addgene, 61592). The nuclease inactivated HiFidCas9 (D10A/H840A mutations), and SpRYdCas9 (D10A/H840A mutations), AsdCas12a (D908A mutation), LbdCas12a (D832A mutation), SadCas9 (D10A and N580A mutations), DpdCas12e (D672A/E769A/D935A mutations), PxdCas12e (D659A/E756A/D922A mutations), dCas12j3 (D413A mutations), and dCas12j2 (D394A mutation) were PCR-amplified using corresponding primer sets designed to engender-specified nuclease-inactivating mutations. Nuclease active and inactivated Cas plasmids were constructed by cloning PCR-amplified fragments into the AfeI and BamHI digested dCas9-dMSK1-P2A-Puro plasmid backbone described previously (Addgene, 165602) via a NEBuilder HiFi DNA Assembly (NEB, E2621). The VPR and KRAB effector domains were amplified from SP-dCas9-VPR (Addgene, 63798) and hUbC-dCas9-KRAB-T2A-Puro (Addgene, 71236), respectively, and cloned into each corresponding BamHI digested dCas plasmid backbone via a NEBuilder HiFi DNA Assembly. The parental CRISPRa testbed plasmids was created by assembling the EcoRI- and MluI-digested dCas9-dMSK1 (Addgene, 63799) together with a PCR-amplified hPGK promoter (Addgene, 63799), a PCR-amplified Blasticidin resistance gene (Addgene, 63799), a PCR-amplified eGFP gene, and a commercially synthesized fragment (gBlock, IDT) harboring NheI and XhoI cut sites upstream of miniCMV. The parental CRISPRi testbed plasmid was created by digesting with the parental CRISPRa testbed plasmid with EcoRI and MluI and then cloning in an EFS promoter (Addgene, 63798) via NEBuilder HiFi DNA Assembly. Each subsequent CRISPRa or CRISPRi testbed plasmid was made by digesting the corresponding parental plasmid with NheI and XhoI and ligating two annealed oligos encoding specified PAM and protospacer sequences. SpCas9-associated gRNAs were cloned into pSPgRNA (Addgene 47108). SaCas9-associated gRNAs were cloned into a pZDonor plasmid containing an SaCas9-gRNA scaffold cassette downstream of hU6 promoter. The gRNA scaffold cassettes for AsCas12a, LbCas12a, Cas12e, Cas12j2, and Cas12j3 were PCR-amplified from pY026 (Addgene, 84741), pY027 (Addgene, 84742), pPP441 (Addgene, 158801), and pPP444 (Addgene, 158802), respectively, and then cloned into NdeI and SacII digested pSPgRNA (Addgene 47108). The gRNAs for each Cas were cloned into the compatible gRNA backbones. All gRNA protospacer sequences are shown in Table S1. Amino acid sequences for Cas constructs are shown in Supporting Notes 1–3.
HEK293T (ATCC and CRL-11268) and HeLa cells (ATCC and CCL-2) were cultured in Dulbecco’s modified Eagle’s medium (Gibco, 31-053-028) supplemented with 10% FBS (Sigma, F2442) and 1% penicillin/streptomycin at 37 °C and 5% CO2. U2OS cells (ATCC, HTB-96) were cultured in a McCoy’s 5A (modified) medium (Gibco, 16-600-082) supplemented with 10% FBS (Sigma, F2442) and 1% penicillin/streptomycin at 37 °C and 5% CO2. Transient transfections were performed in 24-well plates using 375 ng of Cas or dCas (nuclease inactivated) expressing vector and 125 ng of corresponding gRNA vectors. Plasmids were transfected using Lipofectamine 3000 (ThermoFisher, L3000015) following manufacturer’s instructions.
One day before transfection, HEK293T cells were seeded at ∼40% confluency in a 10 cm plate. The next day cells were transfected at ∼80–90% confluency. For each transfection, 10 μg of plasmid containing the vector of interest, 10 μg of pMD2.G (Addgene, 12259), and 15 μg of psPAX2 (Addgene, 12260) were transfected using calcium phosphate. Five hours post-transfection the media was changed. The supernatant was harvested 24 and 48 h post-transfection and filtered with a 0.45 μm PVDF filter (Millipore, SLGVM33RS), and then the virus was concentrated using a Lenti-X Concentrator (Takara, 631232), aliquoted, and stored at −80 °C until use. Lentiviral titers were measured using a Lenti-X qRT-PCR Titration Kit (Takara, 631232).
SDS-page gels were loaded with 25 or 50 μg of total protein and transferred onto a PVDF membrane (Bio-Rad, 1704274) using semi-dry electroblotting (Bio-Rad, 1704150) according to manufacturer’s instructions. Mouse primary α-FLAG antibody (Sigma-Aldrich, F1804) was diluted at 1:1000 in Tris-Buffered Saline with 1% Casein (Bio-Rad, 1610782) and secondary α-mouse HRP-conjugated antibody (Cell Signaling, 7076) was used at a 1:3000 dilution in Tris-Buffered Saline with 1% Casein (Bio-Rad, 1610782). Membranes were incubated in an enhanced chemiluminescence substrate (ECL, Bio-Rad, 1705062). Tubulin was detected with a human α-Tubulin Rhodamine-conjugated antibody (Bio-Rad, 12004165) at a 1:3000 dilution in Tris Buffered Saline With 1% Casein (Bio-Rad, 1610782).
Transfected cells were trypsinized and then washed with phosphate-buffered saline (PBS, Fisher, BP3994) and then, fixed for 12 min in 1.6% formaldehyde (Sigma, F8775-25ML). Fixed cells were then washed with PBS and permeabilized for 15 min with 0.1% Triton X-100 (Sigma, T9284-100ML) in PBS. Permeabilized cells were then washed with PBS and blocked for 30 min with 1% bovine serum albumin (BSA, Fisher, BP9706-100) and 0.1% Tween-20 (Millipore, 655204-100ML) in PBS. Following blocking, cells were incubated with α-FLAG-FITC antibodies (Sigma, F4049-.2MG) diluted in blocking buffer (1% BSA and 0.1% Tween-20 PBS) at a final concentration of 1 μg/mL for 1 h at room temperature. Cells were washed with blocking buffer and analyzed using a Sony SA3800 flow cytometer.
The GFP-PEST HEK293T reporter cell line was generated via a lentiviral integration as described above and previously. Briefly, HEK293T cells were transduced with lentivirus expressing an eGFP-PEST reporter construct under the EFS promoter at a MOI of 10.0. Cells with robust GFP expression were sorted on a MA900 and banked. eGFP HEK293T reporter cells were seeded into 24 well plates and transfected at 60–70% confluency the next day according to the manufacturer’s protocol with lipofectamine 3000 (ThermoFisher, L3000015) using 375 ng of indicated Cas plasmids and 125 ng indicated gRNAs. Cells were analyzed 3 days post-transfection using a SA3800 flow cytometer.
DNA was isolated from the transfected cell 72 h post-transfection using a DNeasy Blood & Tissue kit (Qiagen, 69506). PCR was performed using 50 ng of extracted DNA, primers specific for the targeted EMX1 locus, andQ5 polymerase (NEB, M0491S) following manufacturer’s instructions. PCR products were then purified (Qiagen, 28106) and 100 ng of PCR products were sequenced via Sanger sequencing (Eurofins). Sequencing data was then processed using the TIDE web tool (http://tide.nki.nl). Predicted Sp gRNA nuclease activities using Azimuth 2 algorithm are shown in Table S2.
RNA was isolated from transfected cells using a RNeasy Plus mini kit (Qiagen, 74136) and 1 μg of purified RNA was used as a template for cDNA synthesis (Bio-Rad, 1725038). Real-time quantitative PCR (qPCR) was performed using Luna qPCR Master Mix (NEB, M3003E) and a CFX96 Real-Time PCR Detection System with a C1000 Thermal Cycler (Bio-Rad, 1855195). Baselines were subtracted using the baseline subtraction curve fit analysis mode and thresholds were automatically calculated using the Bio-Rad CFX Manager software version 2.1. Results are expressed as the fold change above mock transfected control cells after normalization to GAPDH expression using the ΔΔCt method. Quantification of plasmid transfection efficiency was performed by creating standard curve serial dilutions of previously reported plasmid (NMS-dCas9-VP64) containing a WPRE. Results were fitted to the standard curve to calculate the number of plasmids transfected, then normalized to GAPDH. All qPCR primers and conditions are listed in Table S3.
The CRISPRa and CRISPRi reporter cell lines were generated using lentiviral integration as previously described. Briefly, HEK293T cells were transduced with lentivirus expressing an eGFP reporter under the miniCMV promoter at an MOI of 10.0 for CRISPRa experiments, or under an EFS promoter at an MOI of 10.0 for CRISPRi experiments, similar to previous designs. Subsets of the populations displaying robust eGFP expression were sorted by selecting approximately 10% of the total average fluorescent population using a Sony MA900 Cell Sorter. HEK293T CRISPRa or CRISPRi reporter cells were seeded into 24 well plates and transfected at 60–70% confluency the next day according to the manufacturer’s protocol with lipofectamine 3000 (ThermoFisher, L3000015) using 375 ng of indicated Cas plasmids and 125 ng associated gRNAs (Table S1). The eGFP intensity was analyzed 3 days post-transfection using a Sony SA3800 flow cytometer and compared to a non-targeting protospacer control.
Data was analyzed using Student’s t-test. Alternative statistical analyses are presented in the source data file, along with all other source data. | true | true | true |
PMC9594874 | Xiaoying Wang,Qian Li,Siyu He,June Bai,Cui Ma,Lixin Zhang,Xiaoyu Guan,Hao Yuan,Yiying Li,Xiangrui Zhu,Jian Mei,Feng Gao,Daling Zhu | LncRNA FENDRR with m6A RNA methylation regulates hypoxia-induced pulmonary artery endothelial cell pyroptosis by mediating DRP1 DNA methylation | 25-10-2022 | lncRNA FENDRR,Pyroptosis,m6A RNA methylation,Dynamin-related protein 1,Pulmonary artery endothelial cells | Background Pyroptosis is a form of programmed cell death involved in the pathophysiological progression of hypoxic pulmonary hypertension (HPH). Emerging evidence suggests that N6-methyladenosine (m6A)-modified transcripts of long noncoding RNAs (lncRNAs) are important regulators that participate in many diseases. However, whether m6A modified transcripts of lncRNAs can regulate pyroptosis in HPH progression remains unexplored. Methods The expression levels of FENDRR in hypoxic pulmonary artery endothelial cells (HPAECs) were detected by using quantitative real-time polymerase chain reaction (qRT-PCR) and fluorescence in situ hybridization (FISH). Western blot, Lactate dehydrogenase (LDH) release assay, Annexin V-FITC/PI double staining, Hoechst 33342/PI fluorescence staining and Caspase-1 activity assay were used to detect the role of FENDRR in HPAEC pyroptosis. The relationship between FENDRR and dynamin-related protein 1 (DRP1) was explored using bioinformatics analysis, Chromatin Isolation by RNA Purification (CHIRP), Electrophoretic mobility shift assay (EMSA) and Methylation-Specific PCR (MSP) assays. RNA immunoprecipitation (RIP) and m6A dot blot were used to detect the m6A modification levels of FENDRR. A hypoxia-induced mouse model of pulmonary hypertension (PH) was used to test preventive effect of conserved fragment TFO2 of FENDRR. Results We found that FENDRR was significantly downregulated in the nucleus of hypoxic HPAECs. FENDRR overexpression inhibited hypoxia-induced HPAEC pyroptosis. Additionally, DRP1 is a downstream target gene of FENDRR, and FENDRR formed an RNA–DNA triplex with the promoter of DRP1, which led to an increase in DRP1 promoter methylation that decreased the transcriptional level of DRP1. Notably, we illustrated that the m6A reader YTHDC1 plays an important role in m6A-modified FENDRR degradation. Additionally, conserved fragment TFO2 of FENDEE overexpression prevented HPH in vivo. Conclusion In summary, our results demonstrated that m6A-induced decay of FENDRR promotes HPAEC pyroptosis by regulating DRP1 promoter methylation and thereby provides a novel potential target for HPH therapy. Supplementary Information The online version contains supplementary material available at 10.1186/s10020-022-00551-z. | LncRNA FENDRR with m6A RNA methylation regulates hypoxia-induced pulmonary artery endothelial cell pyroptosis by mediating DRP1 DNA methylation
Pyroptosis is a form of programmed cell death involved in the pathophysiological progression of hypoxic pulmonary hypertension (HPH). Emerging evidence suggests that N6-methyladenosine (m6A)-modified transcripts of long noncoding RNAs (lncRNAs) are important regulators that participate in many diseases. However, whether m6A modified transcripts of lncRNAs can regulate pyroptosis in HPH progression remains unexplored.
The expression levels of FENDRR in hypoxic pulmonary artery endothelial cells (HPAECs) were detected by using quantitative real-time polymerase chain reaction (qRT-PCR) and fluorescence in situ hybridization (FISH). Western blot, Lactate dehydrogenase (LDH) release assay, Annexin V-FITC/PI double staining, Hoechst 33342/PI fluorescence staining and Caspase-1 activity assay were used to detect the role of FENDRR in HPAEC pyroptosis. The relationship between FENDRR and dynamin-related protein 1 (DRP1) was explored using bioinformatics analysis, Chromatin Isolation by RNA Purification (CHIRP), Electrophoretic mobility shift assay (EMSA) and Methylation-Specific PCR (MSP) assays. RNA immunoprecipitation (RIP) and m6A dot blot were used to detect the m6A modification levels of FENDRR. A hypoxia-induced mouse model of pulmonary hypertension (PH) was used to test preventive effect of conserved fragment TFO2 of FENDRR.
We found that FENDRR was significantly downregulated in the nucleus of hypoxic HPAECs. FENDRR overexpression inhibited hypoxia-induced HPAEC pyroptosis. Additionally, DRP1 is a downstream target gene of FENDRR, and FENDRR formed an RNA–DNA triplex with the promoter of DRP1, which led to an increase in DRP1 promoter methylation that decreased the transcriptional level of DRP1. Notably, we illustrated that the m6A reader YTHDC1 plays an important role in m6A-modified FENDRR degradation. Additionally, conserved fragment TFO2 of FENDEE overexpression prevented HPH in vivo.
In summary, our results demonstrated that m6A-induced decay of FENDRR promotes HPAEC pyroptosis by regulating DRP1 promoter methylation and thereby provides a novel potential target for HPH therapy.
The online version contains supplementary material available at 10.1186/s10020-022-00551-z.
Hypoxic pulmonary hypertension (HPH) is a serious cardiovascular disease characterized by functional and structural changes of in pulmonary vasculature, which leads to increased pulmonary vascular resistance and remodeling, right ventricular hypertrophy, and finally death (Humbert 2019; McGoon et al. 2014; Thompson and Lawrie 2017). Studies have shown that endothelial cells (ECs) are the direct targets of hypoxia and are involved in the pathogenesis process of HPH leading to cell hyperproliferation, inhibition of apoptosis and plexiform intima injuries (Ranchoux et al. 2018; Masri et al. 2007). Therefore, dysfunctional ECs are important players in the imbalance of pulmonary vascular homeostasis and the pathogenesis of HPH. Pyroptosis, a novel form of proinflammatory programmed cell death, is mainly dependent on the activation of caspase-1 or caspase-11 in inflammasomes (Shi et al. 2017; Bergsbaken et al. 2009). Studies have demonstrated that pyroptosis plays an important role in the development of infectious diseases, neurological diseases, atherosclerosis, acute and chronic liver diseases and immune system deficiency diseases (Man et al. 2017; McKenzie et al. 2020; Hoseini et al. 2018; Luan and Ju 2018; Mistry and Kaplan 2017). A recent study provided evidence that pyroptosis is involved in the inflammatory process of human pulmonary artery smooth muscle cells (HPASMCs) in HPH (Zhang et al. 2020). However, the role of pyroptosis in HPAEC, and its potential relationship with HPH as well as the underlying mechanisms are unclear. Long noncoding RNAs (lncRNAs) are a large class of RNA molecules ranging in length from 200 to 100,000 nt and located in the nucleus or cytoplasm (Ponting et al. 2009). A large body of evidence has demonstrated that lncRNAs are engaged in various diseases, including HPH. For example, lncRNA-MEG3 sequesters miR-328-3p, leading to increased expression of IGF1R (Type 1 insulin-like growth factor receptor) and regulating the development of HPH (Xing et al. 2019). However, the reported lncRNAs in HPH are all focused on the cytoplasm of PASMCs, and the specific lncRNAs located in the nucleus, especially in PAECs involved in HPH progression and their related regulatory mechanism, remain largely unknown. LncRNA FOXF1-AS1, also known as FENDRR, with its gene 3099 nt in length, is involved in some types of cancers, including cervical cancer, lung cancer and breast cancer. FENDRR could inhibit cervical cancer proliferation and invasion by targeting miR-15a/b-5p and regulating tubulin alpha1A expression (Zhu et al. 2020). FENDRR suppressed the progression of nonsmall cell lung cancer by regulating the miR-761/TIMP2 (tissue inhibitor of metalloproteinase 2) axis (Zhang et al. 2019). Because HPH has cancer-like phenotypes, such as hyperproliferation and apoptosis resistance, it is possible that FENDRR is involved in HPH pathogenesis. N6-methyladenosine (m6A) RNA modification has been identified to regulate the expression of lncRNAs (Chen et al. 2020; He et al. 2020a).It is one of the most common epitranscriptomic modifications in eukaryotic RNAs and is regulated by m6A "writer" proteins (METTL3, METTL14), "eraser" proteins (FTO, ALKBH5) and "reader" proteins (YTHDC1-3, YTHDF1-3) (Zaccara et al. 2019; Wu et al. 2017). Recent studies have shown that m6A RNA modification plays an important role in all aspects of lncRNA metabolism by regulating the splicing, stability, translocation and translation of transcripts (Dai et al. 2018; Pan 2013). More importantly, m6A modification is reported to be important in the progression of multiple diseases (Jiang et al. 2021a). Although the roles of FENDRR in cancer have been elucidated, the specific mechanism of epigenetic modification regulation in FENDRR remains poorly understood. In the current study, we identified that FENDRR was downregulated in hypoxic HPAEC, and involved in hypoxia-induced pyroptosis of HPAECs. Mechanistically, YTHDC1-mediated m6A modification induced the downregulation of FENDRR, which subsequently promoted hypoxia-induced HPAEC pyroptosis by decreasing the formation of an RNA–DNA triplex in the promoter region of dynamin-related protein 1 (DRP1) to inhibit promoter DNA methylation. Our results reveal a novel regulatory mechanism for hypoxia-induced HPAEC pyroptosis and provide a potential target with therapeutic implications in HPH.
Healthy male C57BL/6J mice with a mean weight of 30 g were obtained from the Experimental Animal Center of Harbin Medical University (Harbin, China). To confirm the role of the functional fragment TFO2 of FENDRR (464–516) in HPH, The TFO2 sequence of FENDRR cloning construction and serotype 5 adenovirus-associated virus (AAV 5) packaging experiment were constructed by HANBIO (Shanghai, China). An aliquot of the vector at 1011 genome equivalents was prepared in 20–30 μL of HBSS and isoflurane anesthesia followed by nasal drops. Mice were randomly divided into five groups as follows: normoxic environment plus control vector group (NOR + NC, n = 20), hypoxic environment plus control vector group (HYP + NC, n = 10), hypoxic environment plus FENDRR TFO2 adenovirus group (HYP + FENDRR TFO2, n = 10), normoxic environment plus FENDRR TFO2 adenovirus group (NOR + FENDRR TFO2, n = 10). Seven days later, mice were assigned to normoxia (Fi,O2 0.21) and hypoxia (Fi,O2 0.10) for seven days as previously described (Zhu et al. 2003). The mice were administered nontargeted control vector or FENDRR TFO2 adenovirus intranasally again and were assigned to normoxia (Fi,O2 0.21) and hypoxia (Fi,O2 0.10) for 14 days. All mice were anaesthetized through an intraperitoneal injection of avertin (200 mg/kg i.p., Sigma-Aldrich, St Louis, USA). For the right ventricular hypertrophy index (ratio of right ventricular free wall weight over the sum of septum plus left ventricular free wall weight: RV/(LV + Sep) calculation, hearts were excised and atria were removed. The RV free wall was dissected, and each cham-ber was weighed.
The right ventricular systolic pressure (RVSP) and echocardiography were measured as previously described (Liu et al. 2020a). The right ventricular systolic pressure (RVSP) was measured with PowerLab monitoring equipment (AD Instruments, Colorado Springs, CO). A 1.2 French Pressure Catheter (Scisense Inc, USA) was inserted into the superior vena cava and finally into the right ventricular vein, and the RVSP was continuously recorded for 20–40 min. Mice were subjected to echocardiography using a Vevo2100 imaging system (VisualSonics Inc., Toronto, Ontario, Canada), pulmonary artery velocity time integral (PAVTI), pulmonary artery acceleration time (PAAT) and left ventricular ejection fraction (LVEF) were obtained from stable images.
Hematoxylin and eosin staining (HE staining) was performed according to the manufacturer’s instructions. In brief, lung tissues of mice were immersed in 4% paraformaldehyde for 48 h. Next, the fixed lung tissues were dehydrated, cleared and embedded in paraffin wax. The lung tissue volume of each block was sampled with equal probability. The paraffin blocks were cut into 5-μm-thick sections and stained with hematoxylin and eosin (HE). In situ hybridization was performed with kits following the manufacturer’s instructions (Boster, Wuhan, China). Digoxigenin-labeled DNA probes complementary to TFO2 of FENDRR were generated using random primer labeling. For each slice stained, 6 high power fields were randomly selected for analysis. The total wall thickness and positive staining area in the vascular walls were quantified by using a color-recognition algorithm in Image-Pro Plus 6.0 software.
HPAECs used in the experiment were purchased from ScienCell Research Laboratories (CA, USA). HPAECs were maintained in endothelial cell medium (ScienCell, 1001, CA, USA) containing 15% fetal bovine serum and 1% penicillin streptomycin at 37 °C, 5% CO2, and 100% relative humidity. Cells under hypoxic conditions were incubated in a Tri-Gas Incubator (Heal Force) with a water-saturated atmosphere comprising 3% O2, 5% CO2 and 91% N2 for 24 h.
Fluorescence-conjugated FENDRR probes were synthesized by RuiBo (Guangzhou, China). FISH experiments were performed using a Fluorescent In Situ Kit (RuiBo Biology, Guanzhou, China) following the manufacturer’s instructions. Briefly, HPAECs were cultured on coverslips and then grown to approximately 60%. After being treated with agents according to the different experimental groups, cells were washed with 1 × PBS, fixed with 4% paraformaldehyde, and permeabilized with 0.3% Triton X-100. Then, the cells were blocked with prehybridization solution at 37 °C for 1 h and incubated with hybridization solution containing FENDRR, 18S and U6 probes overnight at 37 °C in the dark. Finally, 4′,6-diamidino-2-phenylindole (DAPI) was added to stain the nuclei at 37 °C for 10 min. Images were captured with a living cell workstation (AF6000; Leica, Germany).
Cytoplasmic and nuclear RNAs were isolated and purified using a Norgen’s Cytoplasmic & Nuclear RNA Purification Kit (Thorold, ON, Canada) following the manufacturer’s instructions. In brief, 1 × 106 HPAECs were lysed with ice-cold lysis buffer, and cytoplasmic RNA and nuclear RNA were bound to the column. Finally, the mixture was separated for RNA elution analysis.
Protein samples were extracted from HPAECs by using ice-cold lysis buffer and then centrifuged at 13,500 rpm for 15 min at 4 ℃. After centrifugation, the protein concentrations were determined using a Bio-Rad protein assay kit (Bio-Rad Laboratories, Inc., Berkeley, CA, USA). Protein samples (30 µg) were fractionated on 12% SDS-PAGE gels, transferred onto nitrocellulose membranes, and subsequently blocked with 5% nonfat milk at room temperature for 1 h. The membranes were incubated with specific antibodies against NLRP3 (2 µg/mL, bs-10021R, Bioss, Beijing, China), Caspase-1 (1 µg/mL, 22915-1-AP, Proteintech, IL, USA), pro-Caspase-1 (1 µg/mL, ab179515, Abcam, MA, USA), IL-1β (1 µg/mL, 16806-1-AP, Proteintech, IL, USA), DRP1 (1 µg/mL, ab184247, Abcam, MA, USA), YTHDC1 (2 µg/mL, 14392-1-AP, Proteintech, IL, USA), FTO (2 µg/mL, bs-7056R, Bioss, Beijing, China). Bands were sequentially incubated with horseradish peroxidase-labeled secondary antibodies at room temperature for 1 h and enhanced chemiluminescent reagent imaging.
Total RNA was extracted from HPAECs using TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. The concentration and purity of all samples were measured via a NanoDrop 2000 (Thermo Scientific, Wilmington, USA), and cDNA was synthesized with the use of the Superscript first-strand complementary DNA synthesis kit (Invitrogen). Finally, the products were quantified using SYBR Green real-time PCR (Toyobo, Japan) in a Roche LightCycler 480II instrument. The nucleotide sequences of the primers are shown in Table 1.
For the overexpression assay, FENDRR and DRP1 plasmids were constructed using the vector GV219, and empty vector alone was used as a negative control (GeneChem, Shanghai, China). HPAECs were transfected with 3 μg of plasmids using Lipofectamine 2000 reagent following the manufacturer’s instructions. Then, 4–6 h after transfection, the cells were switched to 5% serum-containing medium and cultured under normoxic or hypoxic growth conditions for another 24 h.
The expression of genes was silenced by transfecting HPAECs with small interfering RNAs (siRNAs) or antisense oligonucleotides (ASOs), which were designed and synthesized by GenePharma (Shanghai, China) and RiboBio (Guangzhou, China). HPAECs were transfected with 2 μg of siRNAs or ASO using X-tremeGene siRNA transfection reagents following the manufacturer’s instructions. Six hours after transfection, the cells were switched to 5% serum-containing medium and cultured under normoxic or hypoxic growth conditions for another 24 h. The detailed siRNA sequences are shown in Table 1.
HPAECs were cultured on coverslips until the cell confluence reached 80%. The cells were treated with different agents according to the different experimental groups. Afterward, the cells were stained with 6 μL of Hoechst 33342 solution and 6 μL of PI (propidium iodide) at 4 °C in the dark for 20 min. Images were captured with a living cell workstation (AF6000; Leica, Germany).
HPAECs were plated into 96-well plates at a density of 5000 cells/well, and then cells were treated with different agents according to the different experimental groups. After 24 h of hypoxia, LDH was measured according to the LDH Release Assay Kit instructions (Beyotime Biotechnology, Shanghai, China). Finally, absorbance at 490 nm was recorded.
HPAECs were cultured on coverslips in 12-well plates and then treated with different agents according to the different experimental groups. Prepared cells were washed three times with 1 × PBS and were fixed with 4% paraformaldehyde at 4 °C for 15 min. Then, the cell membrane was permeabilized with 0.3% Triton X-100 for 30 min and blocked with 5% bovine serum for 30 min at room temperature. After that, the cells were incubated with DRP1 (10 µg/mL, ab184247, Abcam, MA, USA), Caspase-1 (6 µg/mL, 22915-1-AP, Proteintech, IL, USA), NLRP3 (10 µg/mL, bs-10021R, Bioss, Beijing, China) and CD31 (5 µg/mL, ab9498 Abcam, MA, USA) antibodies in PBS at 4 °C overnight. After washing three times with 1 × PBS, the cells were subsequently incubated with Cy3-conjugated goat anti-rabbit (A0516, Beyotime, Shanghai, China) and FITC-conjugated goat anti-mouse (A0568, Beyotime, Shanghai, China) antibodies for 2 h at 37 °C in the dark. Cells were then washed with 1 × PBS, DAPI (C1002, Beyotime, Shanghai, China) was added to stain the nuclei at 37 °C for 10 min. Finally, the coverslips were mounted with anti-fade mounting medium (P0126, Beyotime, Shanghai, China) and captured with a living cell workstation. The frozen sections of mouse lung tissues were performed in the same manner.
The RNA immunoprecipitation assay was performed by using an RNA Immunoprecipitation (RIP) Kit (Bes5101, BersinBio, Guangzhou, China) following the manufacturer’s instructions. Briefly, 1 × 107 HPAECs were lysed with RIP lysis buffer. After removing DNA, 20 µL of protein A/G bead-conjugated anti-YTHDC1 antibodies (3 µg, 14392-1-AP, Proteintech, IL, USA) and IgG were added to the samples and incubated overnight at 4 °C. After extracting RNA, the expression of FENDRR was detected by qRT-PCR. For m6A-RNA immunoprecipitation (Me-RIP), an anti-m6A antibody (4 µg A-1801, Epigentek Group Inc., Farmingdale, NY) was used. FENDRR extracted from cell lysates was used to measure the m6A-methylated level of FENDRR.
The interaction between FENDRR and the promoter of DRP1 was determined using chromatin isolation by RNA purification assays according to the instruction manual of the Chromatin Isolation by RNA Purification (ChIRP) Kit (Bes5104, BersinBio, Guangzhou, China). Briefly, 4 × 107 HPAECs were collected and crosslinked with 1% formaldehyde for 20 min at room temperature. Crosslinking was stopped by adding glycine to the cell suspension for 5 min. Then, the cells were lysed with CHIRP lysis buffer and sonicated to obtain DNA fragments of approximately 100–500 bp. Samples were precleared and incubated with FENDRR probes (TFO1, TFO2) at 37 °C for 180 min. Finally, DNA was isolated and subjected to qPCR. The specific biotinylated probes TFO1 and TFO2 were synthesized by GenePharma (Shanghai, China). The sequences are shown in Table 1.
The DRP1 promoter fragment containing the FENDRR binding site was cloned into the GV238 plasmid expressing luciferase (Genepharma, Shanghai, China). HPAECs were cotransfected with the FENDRR expression plasmid and DRP1 plasmid expressing luciferase with Lipofectamine 2000 for 48 h. Then, the luciferase activities were measured by the dual-luciferase reporter assay system (Promega, USA).
The specific biotinylated probe containing the DRP1 TSS fragment and the transcribed FENDRR TFO fragment was synthesized by GenePharma (Shanghai, China). The electrophoretic mobility shift assay (EMSA) was performed by a Chemiluminescent EMSA Kit (GS009, Beyotime, Shanghai, China) following the manufacturer’s instructions. In brief, the biotinylated DRP1 TSS and the synthesized FENDRR TFO2 were reacted in 10 mL of binding reaction buffer at room temperature for 20 min. Then, the sample was added to a 4% nondenatured polyacrylamide gel for electrophoresis purposes and transferred onto a nylon membrane, followed by UV crosslinking at 245 wavelengths. The membrane was incubated with streptavidin-HRP conjugate and enhanced chemiluminescent reagent imaging.
The methylation level of the DRP1 promoter region was detected by a GENMED Universal Gene Methylation Detection Kit (GENMED Scientifics INC.USA). In brief, a genomic DNA extraction kit (K0512, Thermo Scientific, USA) was used to extract genomic DNA. Then, 2 µg of DNA was transformed with GENMED reagents and subsequently subjected to PCR. The PCR programs were as follows: predenaturation at 95 °C for 2 min, 35 cycles of denaturation at 95 °C for 30 s, 57 °C for 90 s, and annealing at 72 °C for 30 s, with the last extension at 72 °C for 5 min. Finally, the PCR products were analyzed by 2% agarose gel electrophoresis and captured with a gel imaging system. The primers for methylation and unmethylation of the DRP1 promoter were shown in Table 1.
The EpiQuik m6A RNA Methylation Quantification Kit (Colorimetric, P-9005, Epigentek Group Inc., Farmingdale, NY) was used to detect global m6A modifications in total HPAEC RNAs following the manufacturer’s instructions. Briefly, 2 µL of NC, 2 µL of PC and 200 ng of RNA were added into strip wells, and the solution was mixed. m6A were detected using capture and detection antibodies. The detected signal was enhanced and then quantified colorimetrically at a wavelength of 450 nm in a spectrophotometer.
Two micrograms of total RNA were deposited on a nylon membrane (FFN10, Beyotime Biotechnology, Shanghai, China), and then the nylon membrane was crosslinked by UV for 3 min. Next, the nylon membrane was stained by using methylene blue. Subsequently, the nylon membrane was blocked for 1 h in blocking buffer, and the membrane was incubated with m6A antibody (1 µg/mL, A-1801, Epigentek Group Inc., Farmingdale, NY) at 4 ℃ overnight. The membrane was incubated with horseradish peroxidase-labeled secondary antibodies at room temperature for 1 h and enhanced chemiluminescent reagent imaging.
Pyroptosis of HPAECs was detected by using an Annexin V-FITC Detection Kit (C1062S, Beyotime Biotechnology, Shanghai, China) according to the manufacturer’s instructions. HPAECs were collected and stained with annexin V-FITC and PI at room temperature for 20 min. Afterward, the samples were analyzed using a BD FACSCalibur Flow Cytometer (BD Biosciences, Bedford, MA).
Total RNA samples were extracted from HPAECs of the normoxia group and hypoxia group, and reverse transcribed into cDNA. Then, real-time PCR was used to measure the mRNA expression levels of FENDRR and DRP1. The correlation of FENDRR and DRP1 was analyzed using the Pearson correlation test of GraphPad Prism 8.0, and P < 0.05 was considered significant.
The caspase-1 activity was detected using a caspase-1 activity assay kit (C1102, Beyotime Biotechnology, Shanghai, China) according to the manufacturer’s instructions. Briefly, 2 × 106 HPAECs were harvested and lysed on ice for 15 min, centrifuged at 16,000×g for 15 min, and the supernatant was mixed with synthetic tetrapeptide Ac-YVAD-pNA and incubated at 37 °C for overnight. Finally, absorbance at 405 nm was recorded. The concentrations of total proteins were measured by a Bradford assay kit (P0006, Beyotime Biotechnology, Shanghai, China) according to the manufacturer’s instructions. The caspase-1 activity was calculated by the standard curve of pNA.
To analyze FENDRR localization, the lncATLAS website (http://lncatlas.crg.eu/) was used. Secondary structure analysis of FENDRR was performed using the RNAfold web server (http://rna.tbi.univie.ac.at/cgibin/RNAWebSuite/RNAfold.cgi). The target proteins prediction of FENDRR was performed through AnnoLnc (http://annolnc.gao-lab.org/index.php) and RNAInter (https://www.rna-society.org/raid/). A Venn diagram was shown using Venny2.1 web server (https://bioinfogp.cnb.csic.es/tools/venny/index.html). KEGG and GO enrichment were analyzed with the DAVID website (https://david.ncifcrf.gov/). The FENDRR TFO sequence and DRP1 promoter TTS sequence were identified with LongTarget (http://lncrna.smu.edu.cn/show/DNATriplex). Space structure docking of FENDRR TFO2 and DRP1 TTS was performed using HNADOCK Server (http://huanglab.phys.hust.edu.cn/hnadock/). CpG islands in the DRP1 gene promoter region were analyzed using Methyl Primer Express (http://www.urogene.org/cgi-bin/methprimer/methprimer.cgi). The N6-methyladenosine (m6A) modification site of FENDRR was predicted using the SRAMP prediction server (http://www.cuilab.cn/sramp).
Statistical analyses were performed using GraphPad Prism Software 8.0 (GraphPad Software Inc.). Data are expressed as mean ± SD. All expression values were checked for normal distribution before statistical. Student's t test was used to compare the data between two groups and one-way ANOVA with Tukey post hoc test was used to compare between multiple groups. For non-normally distributed data, we performed nonparametric analyses such as the Mann–Whitney U test for two groups or Kruskal–Wallis test followed by Dunn post-test for multiple groups. Results with 2-tailed of P < 0.05 were considered statistically significant.
In the NCBI browser, FENDRR was located on human chromosome 16:86474525–86508860 and was 3099 nt in length (Fig. 1a). For quantitative analysis of FENDRR, a specific primer was designed (Additional file 1: Fig. S1a). First, to investigate the significance of FENDRR in hypoxic HPASMCs and HPAECs, we examined the expression of FENDRR by qRT–PCR. The results showed that FENDRR was downregulated in a time-dependent manner in hypoxic HPASMCs and HPAECs, but it decreased 9.28-fold at 24 h in HPAECs than in HPASMCs (Fig. 1b and c). Then, we used the lncATLAS website to predict the subcellular localization of FENDRR, the results of which revealed that FENDRR was mainly localized in the nucleus (Additional file 1: Fig. S1b). The distribution of FENDRR mostly in the nucleus was further confirmed by fluorescence in situ hybridization (FISH) analysis (Fig. 1d). Additionally, cellular fractionation experiments showed that FENDRR was downregulated by 1.84-fold and 2.79-fold in both the cytoplasm and nucleus under hypoxic conditions (Fig. 1e). Finally, the RNAfold Web server was used to analyze the secondary structure of FENDRR, which proved its stability (Additional file 1: Fig. S1c). The above results demonstrated that the expression of FENDRR was downregulated in hypoxia-induced HPAECs and may be a key regulator involved in HPH.
Several studies have shown that hypoxia can induce PASMC pyroptosis (Jiang et al. 2021b; He et al. 2020b), but whether hypoxia affects PAECs pyroptosis is unclear. To confirm the role of hypoxia in pyroptosis of HPAECs, pyroptosis-related proteins were first assessed by Western blot. After 24 h of hypoxia, we found that NLRP3 and Caspase-1 were upregulated by 1.60-fold and 1.75-fold high in hypoxia HPAECs (Fig. 2a). Next, we evaluated the functional role of FENDRR in HPAEC pyroptosis under hypoxic conditions using the FENDRR overexpression plasmid. A sketch map for plasmid construction of FENDRR overexpression is shown in the supplement (Additional file 1: Fig. S2a). The overexpression efficiency was verified by qRT–PCR and FISH (Additional file 1: Fig. S2a and b). Hypoxia-induced upregulation of pyroptosis-related proteins expression was reversed by FENDRR overexpression in HPAECs (Fig. 2b). At the same time, hypoxia increased the positive PI staining, and the effect was inhibited by FENDRR overexpression under the same conditions (Fig. 2c). Immunofluorescence assays indicated that FENDRR overexpression attenuated Caspase-1 expression under hypoxic environments (Fig. 2d). Caspase-1 activity levels also were decreased by FENDRR overexpression in hypoxia HPAECs, the relative values are 1.08 ± 0.16, 2.84 ± 0.77, 1.09 ± 0.27 (Fig. 2e). In addition, FENDRR overexpression inhibited the pyroptotic cell death from 10.3% to 4.1% in HPAECs compared with hypoxic groups. (Fig. 2f). The LDH release assay results indicated that FENDRR overexpression attenuated the increased LDH activity from 18.4% to 10.3% in HPAECs exposed to hypoxia (Fig. 2g). Moreover, the results from western blot assay show that the expression of Caspase-4 and Caspase-11 (nonclassical pyroptosis pathway) was unaffected by FENDRR overexpression in hypoxia HPAECs (Additional file 1: Fig. S2c). These results demonstrate that FENDRR is an important participant in the regulation of hypoxia-induced HPAEC classical pyroptosis pathway.
To further reveal the effects of FENDRR in vitro, FENDRR was knocked down in the nucleus of HPAECs using antisense oligonucleotides (ASOs), and a sketch map of the antisense oligonucleotides (ASOs) of FENDRR is shown in Additional file 1: Fig. S3a. The transfection efficiency was verified by qRT–PCR and FISH (Additional file 1: Fig. S3a and b). FENDRR knockdown promoted expression of pyroptosis-related proteins by at least 1.5-fold and positive PI staining in HPAECs (Additional file 1: Fig. S3c and d). Moreover, pyroptotic cell death in HPAECs increased from 1.3% to 3.9% after transfection of FENDRR ASOs (Additional file 1: Fig. S3e). FENDRR knockdown in HPAECs increased the release of LDH activity from 7.3% to 16.7% and upregulated the fluorescence intensity of Caspase-1 (Additional file 1: Fig. S3f and g). Taken together, these results imply that FENDRR negatively regulates the pyroptosis of HPAECs.
It has been reported that many lncRNAs that interact with proteins are essential to a variety of biological processes (Ferre et al. 2016). Therefore, we used AnnoLnc and RNAInter websites to predict proteins associated with FENDRR. Intersection proteins were evaluated by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis with the DAVID web server, and we found that DNM1L, also known as DRP1, is involved in the NOD-like receptor signaling pathway related to pyroptosis (Fig. 3a). Subsequently, qRT–PCR, Western blot and immunofluorescence assays revealed that upregulation of DRP1 was inhibited by at least 1.5-fold in FENDRR-overexpression HPAECs under hypoxic conditions (Fig. 3b–d). As illustrated by correlation analysis (R2 = 0.5439, P < 0.01), FENDRR expression was negatively correlated with DRP1 expression (Fig. 3e). Moreover, to validate whether FENDRR bound to DRP1 in HPAECs, we performed FISH experiments. The statistical significance of Pearson’s coefficient and the result was 0.42 (Pearson’s coefficient > 0.5 was considered meaningful), suggesting that FENDRR did not colocalize well with DRP1 (Fig. 3f). Therefore, we speculated that FENDRR affects cellular functions in another manner.
It has been reported that lncRNAs can interact with gene promoters through Hoogsteen base pairing to form RNA–DNA triplexes, thereby regulating the expression of target genes (Li et al. 2016). Since FENDRR was highly enriched in the nucleus, it is possible that FENDRR interacts with the DRP1 promoter to regulate its expression. Longtarget was used to predict triplex forming oligonucleotides (TFOs) within the FENDRR RNA, and two TFOs with high scores were verified (Fig. 4a). To verify the results, we performed chromatin isolation by RNA purification (CHIRP) assays to explore whether the two TFOs have the ability to bind to the DRP1 promoter. The results showed that DRP1 promoter region was significantly more enriched by tenfold high in the biotin-labeled TFO2 group than that in the NC and biotin-labeled TFO1 groups (Fig. 4b). In addition, a luciferase assay was performed using DRP1 promoter plasmid, containing the DRP1 promoter binding region (WT) of TFO2 and a mutated region (MUT) inserted downstream of a luciferase reporter. The results revealed that the DRP1 promoter WT group obviously decreased luciferase expression compared to the negative control (NC) group approximately threefold after cotransfection with the FENDRR overexpression plasmid, but the DRP1 promoter MUT group showed no notable changes (Fig. 4c). Then, we further used the HNADOCK Server to predict and analyze the 3D structural docking of the FENDRR and DRP1 promoter, and visualized binding complex structure of FENDRR-DRP1 DNA (Fig. 4d). To rule out the possibility of DNA-RNA heteroduplexe formation, RNase H and RNase A were used to analyze the formation of RNA–DNA triplexes by electrophoretic mobility shift assay (EMSA). We found that treatment with RNase H did not affect the mobility of the RNA–DNA complex, so FENDRR could combine with the DRP1 promoter via RNA–DNA triplex formation (Fig. 4e). As DNA methylation is closely related to the gene transcription process, RNA–DNA triplexes may inhibit target genes through DNA methylation (O'Leary et al. 2015). We suspected that FENDRR might regulate DRP1 transcription in a similar manner. To test this possibility, we found CpG islands in the DRP1 promoter region by using the MethPrimer website (Fig. 4f), suggesting that DNA methylation may exist in DRP1. Accordingly, methylation-specific PCR (MSP) was performed to analyze the methylation level in the promoter region of DRP1 after overexpressing FENDRR. The agarose gel results showed no methylation at the DRP1 promoter region under hypoxia, while methylation was observed in HPAECs overexpressing FENDRR (Fig. 4g). Taken together, these experiments indicated that FENDRR may inhibit DRP1 expression by forming triplexes with the promoter of DRP1 and increasing the methylation status of DRP1.
Previous studies have shown that upregulation of DRP1 is involved in pulmonary vascular remodeling in HPH by controlling metabolic pathways and the proliferation of PASMCs and PAECs (Ryan et al. 2015; Chen et al. 2018; Shen et al. 2015). However, whether DRP1 regulates hypoxia-induced HPAEC pyroptosis has not been reported. We transfected DRP1 siRNA into HPAECs to knock down DRP1 protein expression, interference efficiency was 70% (Additional file 1: Fig. S4a). The increased expression of the proteins NLRP3, Caspase-1, pro-Caspase-1 and IL-1β under hypoxia was reduced by DRP1 siRNA (Fig. 5a). We further observed that DRP1 siRNA decreased the level of LDH activity from 15.4% to 8.0% under hypoxic conditions (Fig. 5b). Caspase-1 activity levels also were decreased by DRP1 siRNA in hypoxia HPAECs, the relative values are 0.93 ± 0.17, 1.76 ± 0.13, 1.01 ± 0.06 (Fig. 5c). Moreover, PI staining and fluorescence intensity of Caspase-1 was observed to be increased under hypoxia exposure and reversed by silencing the DRP1 gene (Fig. 5d and e). Finally, we performed rescue experiments on cell pyroptosis to determine whether DRP1 is involved in FENDRR-mediated HPAEC pyroptosis. DRP1 protein expression was significantly upregulated by fourfold high after transfection of the DRP1 overexpression plasmid into HPAECs (Additional file 1: Fig. S4b). Overexpression of DRP1 partially restored the increased levels of pyroptosis-related proteins inhibited by FENDRR overexpression under hypoxia (Fig. 5f). LDH release assay demonstrated that the FENDRR overexpression-induced decrease LDH release was rescued by DRP1 overexpression under hypoxic conditions, LDH release value are 7.98%, 16.83%, 9.43%, 16.61% (Fig. 5g). In addition, this FENDRR-mediated attenuation of hypoxia-induced PAEC Caspase-1 activity levels could be rescued by DRP1 overexpression, Caspase-1 activity relative values are 0.90 ± 0.09, 2.11 ± 0.46, 0.86 ± 0.19, 1.93 ± 0.46 (Fig. 5h). Similar results were observed in PI staining and immunofluorescence of Caspase-1 assays (Fig. 5i and j). These data indicated that DRP1, as a downstream target gene of FENDRR, is involved in FENDRR-mediated HPAEC pyroptosis under hypoxia.
Recent studies have suggested that m6A RNA modification is a novel mediator of pathological changes in PH by regulating diverse transcripts (Hu et al. 2021; Xu et al. 2021). Therefore, we attempted to explore how m6A affects the FENDRR transcript under hypoxic conditions. To address this issue, we verified the m6A modification of FENDRR by using the SRAMP prediction server, and found eight high confidence m6A modification sites of FENDRR (Fig. 6a). The FENDRR target proteins predicted in Fig. 4a contained several m6A modification enzymes, including FTO and YTHDC1 located in the nucleus. Next, YTHDC1 and FTO expression in HPAECs was evaluated using Western blot. The results showed that the expression of YTHDC1 was enriched by 1.4-fold high in HPAECs under hypoxic conditions; FTO did not change significantly. Therefore, we chose YTHDC1 for subsequent analyses (Fig. 6b). We transfected YTHDC1 siRNA into HPAECs to silence YTHDC1, and si-1 had the most obvious effects, interference efficiency was 50% (Additional file 1: Fig. S5a and b). To clarify whether YTHDC1 affects FENDRR expression under hypoxia exposure, HPAECs were transfected with YTHDC1 siRNA. qRT–PCR showed decreased expression of FENDRR under hypoxia, which was reversed by YTHDC1 siRNA transfection, the relative values are 1.00 ± 0.00, 0.56 ± 0.19, 10.2 ± 8.88 (Fig. 6c). Similar results were observed in the FISH experiment, suggesting that the downregulation of FENDRR was mediated by YTHDC1 in hypoxia (Fig. 6d). In addition, MeRIP-PCR results showed that the m6A level of FENDRR was increased by 1.6-fold high in YTHDC1-silenced HPAECs compared with that in the NC group under hypoxia (Fig. 6e). We further detected global m6A RNA modification in HPAECs using the EpiQuik™ m6A RNA Methylation Quantification Kit (Colorimetric, P-9005-48, USA), the results showed that the global m6A RNA modification was increased by 3.7-fold high in YTHDC1-silenced HPAECs compared with that in the NC group under hypoxia (Fig. 6f). Similar trends in dot blot were found in HPAECs after YTHDC1 siRNA treatment (Fig. 6g). Studies have pointed out that YTHDC1 can regulate RNA stability (Liang 2021), so the half-life of FENDRR was tested by incubation with the transcription inhibitor actinomycin D on YTHDC1-silenced HPAECs, and RNA was obtained at different time points. Indeed, silencing YTHDC1 significantly prolonged the lifetime of FENDRR under hypoxic conditions (Fig. 6h). Therefore, YTHDC1 decreased FENDRR stability, indicating the leading role of YTHDC1 in the m6A-mediated degradation process of FENDRR. To test the interaction between FENDRR and YTHDC1, we performed a RIP assay, and the result suggested that YTHDC1 was significantly enriched by tenfold in FENDRR RNA to form an m6A modification complex (Fig. 6i). Fluorescence colocalization assays revealed that FFNEDRR and YTHDC1 were colocalized in the nuclei of HPAEC, the statistical significance of Pearson’s coefficient and the result was 0.72 (Pearson’s coefficient > 0.5 was considered meaningful) (Fig. 6j). Moreover, Western blot indicated that YTHDC1 siRNA reduced the level of DRP1 compared to hypoxia, the relative values are 1.00 ± 0.00, 1.63 ± 0.49, 0.89 ± 0.21 (Fig. 6k). In conclusion, these results showed that YTHDC1 selectively binds to m6A-modified FENDRR and modulates its degradation in an m6A-dependent manner, providing a novel view for the dysregulation of FENDRR in HPH progression.
Although we did not find the mouse homolog of FENDRR, the functional fragment TFO2 of FENDRR (464–516) showed higher conservation of the sequence among humans and mice (Additional file 1: Fig. S6a). To evaluate the function of the conserved TFO2 sequences of FENDRR in the HPH mouse model, adenoviruses with NC or TFO2 sequences were used to treat mice via dropwise intranasal instillation (Fig. 7a). Subsequently, TFO2 overexpression was verified by in situ hybridization methods (Additional file 1: Fig. S6b). Next, we explored whether TFO2 overexpression prevented hypoxia-induced PH in vivo. We characterized the mice in detail, including right ventricular systolic pressure (RVSP), RV/left ventricular (LV) + Septum weight ratio, hemodynamics, cardiac function, and vascular remodeling. The results illustrated that overexpression of the TFO2 fragment inhibited hypoxia-induced RVSP and RV/(LV + S) (Fig. 7b and c). Moreover, we found that TFO2 overexpression reversed hypoxia-induced pulmonary vascular remodeling by HE staining (Fig. 7d). Echocardiographic analysis showed that the pulmonary artery acceleration time (PAAT) and pulmonary arterial velocity time integral (PAVTI) were significantly decreased, and the right ventricle internal diameter (RVID) was increased under hypoxia in mice, while the effect was prevented by TFO2 overexpression. Importantly, the left ventricular ejection fraction (LVEF) was not changed in mice with or without TFO2 overexpression (Fig. 7e). In addition, we examined the effect of TFO2 overexpression on pyroptosis in lung tissues, and mRNA expression levels of pyroptosis-related genes were significantly lower in TFO2 overexpressing mouse lung tissues than in NC-treated mouse lung tissues after hypoxia exposure (Fig. 7f). Immunofluorescence staining for NLRP3 revealed that the fluorescence activity in TFO2 overexpressing mouse lung tissues was decreased compared with that in NC-treated mouse lung tissues after hypoxia exposure (Fig. 7g). Similarly, DRP1-positive staining was inhibited by TFO2 overexpressing under hypoxic conditions (Additional file 1: Fig. S6c). Western blot also indicated that TFO2 overexpressing reduced the protein level of DRP1 compared to hypoxia, the relative values are 0.77 ± 0.24, 1.60 ± 0.11, 0.88 ± 0.32 (Additional file 1: Fig. S6d). To verify the effect of TFO2 adenovirus on PH under normoxic conditions, adenoviruses with NC or TFO2 sequences were used to treat mice via dropwise intranasal instillation (Additional file 1: Fig. S7a). Subsequently, TFO2 overexpression was verified by in situ hybridization methods, TFO2-positive staining was twofold higher than the NOR + NC group (Additional file 1: Fig. S7b). Both the RVSP, RV/left ventricular (LV) + Septum weight ratio and vascular remodeling were unchanged in NOR + NC group and NOR + FENDRR TFO2 group (Additional file 1: Fig. S7c–e). Consistent with these results, echocardiographic analysis showed that the PAAT, PAVTA and LVEF were not changed in mice with or without TFO2 overexpression at baseline conditions (Additional file 1: Fig. S7f). These results confirmed that the conserved TFO2 sequence of FENDRR might regulate the pathology of HPH through pyroptosis in vivo.
In the current study, we proved that FENDRR was localized in the nucleus of PAECs and downregulated in response to hypoxia. Our results support that FENDRR plays a key role in hypoxia-induced HPAECs pyroptosis via regulation of the downstream target DRP1. Mechanistically, FENDRR formed an RNA–DNA triplex within the promoter of DRP1, leading to decreased transcription of DRP1 by promoting DRP1 promoter methylation. More importantly, we showed a new mechanism of FENDRR degradation through binding to the m6A “reader” YTHDC1. In addition, the fragment of the TFO2 sequence (464–516) of FENDRR was the pivotal functional domain that interacted with the DRP1 promoter. The conserved fragment of TFO2 of FENDRR might reverse pyroptosis in an HPH mouse model. These findings implicate FENDRR in PAEC pyroptosis induced by hypoxia, which is a novel mediator in HPH. Increasing evidence suggests that pyroptosis is implicated in cardiovascular diseases, including HPH. It has been reported that PASMC pyroptosis is mediated by circCalm4 and Krüppel zinc finger protein GLI1 in HPH (Jiang et al. 2021b; He et al. 2020b). Moreover, the activation of pyroptosis in PASMCs, which is related to pulmonary fibrosis induced by hypoxia, was alleviated by treatment with Caspase-1 inhibitors (Zhang et al. 2020). However, whether pyroptosis of PAECs is involved in HPH remains largely unknown. In this study, our findings demonstrated that the pyroptosis-related markers NLRP3, Caspase-1 pro-caspase-1 and IL-1β were highly upregulated in hypoxic PAECs. In addition, PI staining and LDH activity increased in PAECs under hypoxic conditions. Therefore, our results provide solid evidence that pyroptosis can occur in PAECs treated with hypoxia, and play a potential role in the development of HPH. Understandably, one of the mechanisms of lncRNAs in the regulation of their biological functions occurs through binding to DNA or proteins in the nucleus to regulate gene transcription and splicing levels (Sun et al. 2018). In the present study, we used bioinformatics analysis to predict proteins with potential sites for binding to FENDRR, and overexpression of FENDRR decreased the mRNA and protein level of DRP1 upon hypoxia exposure, suggesting that DRP1 expression is regulated by FENDRR in PAECs. The possibility of FENDRR and DRP1 protein interaction was ruled out, since the binding tendency between them, as shown by fluorescence colocalization, was very low. We further identified that FENDRR can interact with the DRP1 promoter by performing CHIRP assays and EMSAs. In addition, we confirmed that FENDRR mediated PAEC pyroptosis via DRP1. Therefore, we provide an effective mechanism by which nuclear FENDRR can form RNA–DNA triplexes with the DRP1 promoter, and downregulate the expression of DRP1, thereby inhibiting the occurrence of PAEC pyroptosis. Interestingly, some studies have demonstrated that FENDRR in the cytoplasm participates in the progression of cancer through the ceRNA mechanism (Yu et al. 2019; Cheng et al. 2020), suggesting that FENDRR may play a different role according to its subcellular compartment environment. Recent studies have demonstrated that RNA–DNA triplexes regulate target gene transcription activity by influencing CpG island methylation or transcription factors (Li et al. 2019). In our study, we found that FENDRR forms an RNA–DNA triplex with the DRP1 promoter to enhance the methylation status of the DRP1 promoter CpG island, which inhibits the transcription of DRP1. From the perspective of nuclear FENDRR regulation, we verified that the RNA–DNA triplex structure affects gene transcription through DNA methylation as an epigenetic mechanism underlying the regulatory role of FENDRR in PAEC pyroptosis induced by hypoxia. DPR1 was originally known as a key fission protein in mitochondrial dynamics, and is involved in pathological proliferation and apoptosis resistance of pulmonary vasculature in HPH (Chen et al. 2019; Zhang et al. 2016). Some studies have shown that DRP1 can regulate the occurrence of pyroptosis by influencing mitochondrial homeostasis (Zou et al. 2020). Our present study revealed that DRP1, which regulates hypoxia-induced PAEC pyroptosis, was reversed by FENDRR overexpression in hypoxia-exposed PAECs. More importantly, we observed that overexpression of FENDRR cannot inhibit the generation of mitochondrial reactive oxygen species (ROS) in PAECs caused by hypoxia (Additional file 1: Fig. S8a). Therefore, FENDRR decreased hypoxia-induced PAEC pyroptosis via DRP1 without relying on the regulation of mitochondrial function, suggesting that DRP1 not only has important functions in the mitochondria but also plays a regulatory role as a cytokine in the cytoplasm. N6-methyladenosine (m6A) is the most abundant chemical RNA modification in mRNA and noncoding RNA, and regulates multiple biological processes, such as cell proliferation, migration and tumorigenesis (Liu et al. 2020b). Emerging studies have shown that m6A methylation modification can occur in HPH (Xu et al. 2021). However, there are no studies on the function of m6A in lncRNAs associated with HPH. In this study, we revealed that the m6A “reader” YTHDC1 might negatively regulate FENDRR stability by directly combining with FENDRR. YTHDC1 siRNA significantly prolonged the decay rate of FENDRR as a result of the accumulation of m6A modifications in FENDRR in hypoxic PAECs. These results suggested that YTHDC1 targeted the m6A sites of FENDRR and subsequently decreased FENDRR expression. This is a novel finding regarding the upstream regulatory mechanism of FENDRR under hypoxic conditions.
To conclude, the present study uncovered the involvement of m6A “reader” YTHDC1-medited nuclear FENDRR in hypoxia-induced PAEC pyroptosis by forming RNA–DNA triplex with DRP1 promoter to promote its methylation at CpG islands, along with changes in transcriptional activity, inhibiting the expression of DRP1 (Fig. 8). Our results provide new clues for the study of the molecular regulatory mechanism of pyroptosis in PAECs and may provide potential therapeutic targets in HPH.
Additional file 1: Fig. S1. Specific primer of FENDRR were designed by NCBI; subcellular localization of FENDRR predicted by lncATLAS website and secondary structure of FENDRR. Fig. S2. Overexpression efficiency of FENDRR; the protein levels of Caspase-4 and Caspase-11. Fig. S3. FENDRR ASO enhances cell pyroptosis in HPAECs. Fig. S4. Interference efficiency and overexpression efficiency of DRP1. Fig. S5. Interference efficiency of YTHDC1. Fig. S6. Conservative analysis of the functional fragment TFO2 of FENDRR (464–516); in situ hybridization of the functional fragment TFO2 of FENDRR (464–516); the expression levels of DRP1 were detected by immunofluorescence and western blotting. Fig. S7. Overexpression conserved sequence TFO2 adenovirus of FENDRR in vivo does not affect the development of PH under normoxic conditions. Fig. S8. Mitochondrial superoxide indicator (Mito-SOX Red) was used to detect the mitochondrial-derived ROS production. | true | true | true |
|
PMC9595653 | Janvie Manhas,Lov Raj Lohani,Ashikh Seethy,Uma Kumar,Shivanand Gamanagatti,Sudip Sen | Case report: Characterization of a rare pathogenic variant associated with loss of COL3A1 expression in vascular Ehlers Danlos syndrome | 11-10-2022 | vascular Ehlers Danlos syndrome,COL3A1 pathogenic variant,clinical fibroblast testing,exome sequencing,stroke,hepatic artery dissection | The vascular subtype of Ehlers Danlos Syndrome (vEDS) is a rare connective tissue disorder characterized by spontaneous arterial, bowel or organ rupture. The diagnosis of vEDS is established in a proband by identification of a heterozygous pathogenic variant in the alpha-1 gene of type III collagen (COL3A1) by molecular analysis. In this report, we present a case of vEDS with life threatening, spontaneous arterial dissections in association with an uncharacterized rare variant of COL3A1, exon19:c.1340G > A. Primary culture of patient skin fibroblasts followed by immunofluorescence revealed a complete absence of COL3A1 protein expression as well as altered morphology. Electron microscopy of the cultured fibroblasts showed abnormal vacuoles in the cytoplasm suggestive of a secretory defect. In this study, we have performed functional characterization of the COL3A1 exon19:c.1340G > A variant for the first time and this may now be classified as likely pathogenic in vEDS. | Case report: Characterization of a rare pathogenic variant associated with loss of COL3A1 expression in vascular Ehlers Danlos syndrome
The vascular subtype of Ehlers Danlos Syndrome (vEDS) is a rare connective tissue disorder characterized by spontaneous arterial, bowel or organ rupture. The diagnosis of vEDS is established in a proband by identification of a heterozygous pathogenic variant in the alpha-1 gene of type III collagen (COL3A1) by molecular analysis. In this report, we present a case of vEDS with life threatening, spontaneous arterial dissections in association with an uncharacterized rare variant of COL3A1, exon19:c.1340G > A. Primary culture of patient skin fibroblasts followed by immunofluorescence revealed a complete absence of COL3A1 protein expression as well as altered morphology. Electron microscopy of the cultured fibroblasts showed abnormal vacuoles in the cytoplasm suggestive of a secretory defect. In this study, we have performed functional characterization of the COL3A1 exon19:c.1340G > A variant for the first time and this may now be classified as likely pathogenic in vEDS.
Vascular Ehlers Danlos Syndrome (vEDS; OMIM130050) is a rare connective tissue disorder caused by mutations in the alpha-1 chain of type III collagen (COL3A1) polypeptide leading to serious risk of arterial or organ rupture and premature death (1). It is dominantly inherited in 50% cases and the other 50% present with de novo pathogenic, somatic variants in COL3A1 (2). The most common COL3A1 pathogenic variant is a heterozygous missense substitution for glycine in the (Gly-X-Y) repeating sequence of collagen triple helix which disrupts the assembly of type III homotrimeric collagen (3). This leads to defective type III collagen synthesis and assembly, manifesting as loss of mechanical strength in arteries and other hollow organs. The natural course of vEDS and associated clinical phenotype of patients are both reported to be influenced by the type of COL3A1 variant (3, 4). Therefore, establishing a diagnosis of vEDS should ideally always include molecular genetic testing for COL3A1 pathogenic variants and biochemical analysis to ascertain the pathogenicity of unreported variants (5). Due to limited genetic testing and unavailability of clinical fibroblast culture testing, the prevalence and burden of vEDS in India is underestimated and underreported. vEDS is defined based on one major and several minor diagnostic criteria (1) which highlight the plethora of different physical signs that may constitute the clinical phenotype and cause confusion during diagnosis. Considering the unpredictable, life threatening complications of vEDS with a low median survival of around 48 years (2), it becomes imperative to employ genetic testing early during the management of a suspected vEDS patient especially when the phenotype is indistinguishable from other inherited connective tissue disorders.
A 48-year-old male with no past history of any relevant disease presented with the sudden onset of severe headache followed by right sided hemiparesis. On examination the patient had a thin face with prominent nose and lobeless ears (Figure 1A) along with fragile and thin skin over the extremities suggestive of Acrogeria (Figure 1B). Presence of varicose veins with some ecchymotic patches were observed over the lower limbs (Figure 1C). Chest veins were also found to be prominent (Figure 1A). Hypermobility of small joints of the hands (Figure 1D) and clubbing of feet were observed (Figure 1B). However, hyper elasticity of skin and marfanoid features were not present. Bilateral pitting edema was present in the lower limbs. All other vital parameters were within normal range. Digital Subtraction Angiography (DSA) was suggestive of dissection with sub occlusive narrowing in the left internal carotid artery in the skull base region (Figures 1E,F). Patient was started on anticoagulation therapy with vitamin K antagonist (VKA) and 5 mg of warfarin, dose titrated to maintain a target INR (International Normalized Ratio) of 2–3. Hemiparesis gradually subsided and warfarin was stopped. Patient was asymptomatic for 1 year followed by the sudden onset of upper abdominal pain. Computed tomography (CT) angiography showed aneurysmal dilatation of left hepatic artery (Figures 1G,H) along with multiple arterial dissections involving left renal artery (Figures 1I,J) and multiple infarcts in right renal artery and proximal right common iliac artery (Figures 1K,L). The patient also developed elevated blood pressure which was a new symptom and was managed using anti-hypertensives. A rise in the level of the inflammatory marker, C-reactive protein was observed in serum. Nerve conduction study and nerve biopsy were suggestive of axonal and demyelinating neuropathy in both lower legs. Blood tests for vasculitis markers including antinuclear antibody, perinuclear and cytoplasmic antineutrophil cytoplasmic antibodies (myeloperoxidase and proteinase) were found to be negative. The patient was negative for HIV, hepatitis B and C. Within a few days our patient presented with another episode of acute abdominal pain with bilateral subcostal tenderness. He was diagnosed with hemoperitoneum due to Hepatic Artery rupture with ischemic liver changes, for which aneurysmal coiling was performed. There was no history of weight loss, fever, hemoptysis or melena. Laboratory test results are shown in Supplementary Table 1.
DNA was isolated from the peripheral blood sample of the patient and sent for whole exome sequencing (WES) to identify the cause of the disease. A total of 34058 variants were called after filtering for the depth of sequencing (minimum 10), of which 9257 were exonic variants, including single nucleotide variants and short indels. Exclusion of synonymous variants and filtering for minor allele frequency of <5% retained 854 variants. Once the non-pathogenic variants were removed using SIFT, Polyphen2-HDIV and Mutation Assessor predictions, 548 variants remained (Supplementary Data 1). After inclusion of 28 variants which affected splicing, there were 576 pathogenic variants across 392 different genes, of which 154 were found to be associated with a phenotype in OMIM. On further evaluation based on the patient’s phenotype of ‘blood vessel abnormalities AND hypermobility’ in OMIM, COL3A1 exon19:c.1340G > A [chr2:g.188994587G > A,NM_000090.3:c.1340G>A, NM_000090.3(COL3A1_i001):p.(Gly447Asp), all descriptions are based on hg38] variant leading to the substitution of Aspartic acid for Glycine at codon 447 (p.G447D) in type III collagen α1 polypeptide was identified. The variant attributes according to the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) guidelines (6) for the interpretation of sequence variants were PM2, PP2, and PP3. Visualization of this variant in Integrative Genome Viewer (IGV) suggested heterozygosity (Supplementary Data 1 in Supplementary Figure 1). This variant has been previously found to be associated with vascular subtype of Ehlers Danlos syndrome in one case report (7). Since this variant has not been reported in the 1000 genomes (8), Exome Aggregation Consortium (9) (ExAC) and other databases (10) and no functional data was available regarding this variant, we further proceeded to analyze the pathogenic basis of this variant using molecular techniques.
Patient and control fibroblasts were cultured from skin biopsy (11) (Figures 2A,B) and analysis of protein expression of COL3A1 was done using immunofluorescence. There was a complete absence of COL3A1 production in the patient fibroblasts compared to the control fibroblasts (Figures 2C,D). The morphology of patient fibroblasts was also observed to be altered as compared to the control fibroblasts. Fibroblasts derived from skin biopsy of patient with exon19:c. 1340G > A variant were larger in size as compared to the control fibroblasts (Figures 2C,D). It has been previously reported that dermal fibroblasts from vEDS have defective collagen biosynthesis/processing, endoplasmic reticulum homeostasis/protein folding, disorganized ECM interactions and may acquire a myofibroblast like phenotype (12). In order to investigate the altered morphology, we performed a Transmission electron microscopy (TEM) analysis. TEM has previously been suggested to be a useful addition to the repertoire of diagnostic tests in vEDS (13). Previous studies have reported an inefficient secretion of mutant collagen molecules which can lead to altered structural anatomy of the endoplasmic reticulum and unfolded protein response (14). We observed lamellar vacuolar bodies in the cytoplasm of vEDS fibroblasts which were not present in control fibroblasts (Figures 2E,F). On in-depth observational analysis, these bodies appeared to be “Amphisomes,” organelles of the autophagy pathway resulting from fusion of endosomes and autophagosomes (15). This may explain the altered morphology and size of vEDS fibroblasts.
Genomic DNA was also extracted from cultured dermal fibroblasts from patient and age-matched control and the COL3A1 exon19:c.1340G > A variant obtained from Whole exome sequencing was confirmed by Sanger sequencing (Figure 2G). All methodology provided in Supplementary Data 2.
Diagnosis of vEDS based on clinical diagnostic criteria alone is often difficult due to significant overlap in clinical presentation with other connective tissue disorders and arteriopathies. vEDS diagnosis should be suspected in individuals presenting with spontaneous ruptures of arteries, uterus or bowel at a young age. Careful interpretation of genetic testing results and variant assessment according to the ACMG/AMP guidelines (6) is essential to confirm the diagnosis, as all COL3A1 variants are not pathogenic (16). Additionally, clinical fibroblast testing and functional analysis can help to accurately report pathogenicity of novel sequence variants.
About 705 unique likely pathogenic/pathogenic COL3A1 variants have been reported, which are listed in the Leiden Open Variation Database (see https://databases.lovd.nl/shared/genes/COL3A1). Missense mutation involving substitution of a glycine amino acid in the triple helix is the most common type of mutation reported in the database. The second most common type of mutation is a COL3A1 RNA splicing mutation. Large deletions or insertions in COL3A1 are relatively uncommon in vEDS. Most patients with vEDS develop major complications before the age of 30 years. Autosomal dominant inheritance of vEDS shows a penetrance approaching 100% but may also vary according to the age of presentation. Haploinsufficient patients have a lower penetrance of ∼50% and were observed to survive 10–15 years longer than individuals harboring RNA splicing or glycine mutations (17). Interestingly, a missense mutation with substitution of a valine or an aspartic acid for a glycine was reported to have a poorer prognosis than a substitution of a serine (17). Although the phenotype of patients with amino acid substitutions in COL3A1 may partially overlap with of some haploinsufficient patients (18), there is considerable evidence for genotype-phenotype correlation (3, 16, 17) in vEDS which may be critical for screening and lifelong management of the affected individual and family members. Our patient presented with stroke along with cranial and abdominal visceral arteries’ dissections and rupture. Patient was hypertensive, had an elevated C-reactive protein (CRP) and neuropathy which lead to an initial suspicion of Polyarteritis Nodosa (PAN) (19). However, he did not show constitutional symptoms of fever, fatigue, weight loss, muscle/joint pains, radiological findings or other features defined by American College of Rheumatology classification criteria for PAN. The clinical presentation also favored the differential diagnosis of vEDS, but the presenting age of 48 years was not consistent with this diagnosis. The presence of multiple clinical findings namely acrogeria, progeria, lobeless ear, hypermobile small joints of hands, thin fragile skin, foot clubbing and multiple arterial dissections favored the diagnosis of vEDS. Patient had renal infarcts following renal artery dissection which could have contributed to the rise in blood pressure. It has been reported that delayed clearance of thrombi as well as the vascular insult to arteries in vEDS may lead to a rise in inflammatory markers (20), as observed in this case. Presence of axonal polyneuropathy has also been reported before in vEDS (21, 22). As there was no family history of EDS or related syndromes, in order to establish the diagnosis in a proband it was imperative to perform molecular genetic testing to identify and characterize this variant (2). Corticosteroids are used in PAN to reduce inflammation but they have been reported to trigger arterial dissections by elevating blood pressure and increasing blood vessel fragility by its inhibitory effect on collagen formation and connective tissue strength (23). Therefore, the patient was managed symptomatically till the genetic testing results were obtained. Whole exome sequencing revealed a rare, uncharacterized variant of COL3A1 gene, exon19:c.1340G > A variant. A molecular diagnostic study was carried out to confirm if this variant was responsible for the vEDS phenotype. Dermal Fibroblast culture was established using skin biopsy from the patient and age matched control and COL3A1 protein expression was studied using immunofluorescence. The COL3A1 expression was observed to be absent in the exon19:c.1340G > A variant (Figures 2C,D). This suggests the expression of a mutant collagen with a dominant negative effect leading to a possible degradation after binding with the wild type collagen molecule. The presence of the variant was further confirmed by Sanger sequencing of the PCR amplicon obtained by using DNA extracted from cultured fibroblasts as the template and primers flanking the site of the variant (Figure 2G). We also observed 2 more non-synonymous variants of COL3A1 (exon 30:c.2092G > A and exon 50:c.4059T > G) during the data analysis of the WES. However, based on their high minor allele frequencies, these were found to be non-deleterious (Table 1). Transmission Electron Microscopy (TEM) studies of the dermal fibroblasts suggested that exon19:c.1340G > A variant is associated with altered endoplasmic reticular morphology in vEDS which may contribute to the disease mechanism. According to the ACMG/AMP standards and guidelines for interpretation of sequence variants (6), the following COL3A1 variant criteria may be used to classify exon19:c. 1340G > A variant as pathogenic variant: PS3 → functional studies supporting damaged or altered gene product, absence of COL3A1 protein expression on immunofluorescence studies on skin fibroblasts and altered morphology of fibroblasts on electron microscopy studies. PM2 → variant not reported before in Exome sequencing project 1000 genomes and ExAC. PP2 → missense z-scores of COL3A1 gene > 3.09, which was regarded as intolerant to missense variants. PP3 → SIFT (Deleterious), PolyPhen (Damaging) and a CADD-Phred score of 26.1 (pathogenic) support a deleterious effect on the gene or gene product. Using the ACMG/AMP rules for combining the above criteria (6), COL3A1 exon19:c.1340G > A variant classification was found to be likely pathogenic (II) (Table 1).
Diagnosis and management of a rare, life threatening vascular disease condition in absence of any curative therapy is a difficult conundrum for both the patient and the treating physician. Besides symptomatic treatment and regular follow up, psychological therapy and family support is required to accept and live with such a condition. Clinical evaluation of vEDS patients includes regular blood pressure monitoring, non-invasive arterial screening to detect dissections and dilatations such as ultrasonography, computed tomography or magnetic resonance imaging (1, 2). Any invasive procedures like colonoscopy, angiograms or surgeries should be carefully assessed for risk and if necessary should be performed by surgeons experienced and familiar with the enhanced caution required due to tissue fragility and its associated complications (24). After the confirmation of COL3A1 variant by exome sequencing and biochemical analysis, diagnosis was confirmed as vEDS. Patient was explained about the disease and its prognosis and referred for genetic counseling. It has been reported that Celiprolol improves the biomechanical integrity of the aorta (25) and may be an option for the prevention of complications in vEDS (26). Due to lack of Celiprolol and any other approved therapies in India and many other countries, there are not many options available for treating vEDS patient besides symptomatic and supportive therapy. As research advances further in this field, gene therapy may be a promising option for the treatment of vEDS in the future. However, addition of the defective gene is not applicable for dominant diseases such as vEDS where the defective procollagen gene, COL3A, is a homotrimer of three identical α1 chains. As most of the patients present with a heterozygous mutation, there is a structural defect in half of the synthesized collagen α1 fibrils which results in 1/8th normal and 7/8th abnormal COL3A trimers. Interesting approaches such as siRNA mediated inhibition of the mutated allele and strategies to enhance transcriptional activation of the normal allele are being tested in vivo for phenotypic correction of vEDS (27, 28). A recent study has identified PLC/IP3/PKC/ERK pathway (phospholipase C/inositol 1,4,5-triphosphate/protein kinase C/extracellular signal-regulated kinase) as major drivers of vascular pathology in vEDS based on preclinical models (29). AR101 (enzastaurin) is an orally active, small molecule, serine/threonine kinase inhibitor of the PLC/IP3/PKC/ERK pathways which has been previously studied in over 40 human trials including a range of cancers. PREVEnt trial, which will assess the efficacy of enzastaurin in preventing cardiac or arterial events in patients with vEDS confirmed with COL3A1 gene mutations is scheduled to begin next year and if successful may provide us with a new treatment for the life-threatening complications of vEDS.
We evaluated a rare COL3A1 variant in vEDS using the ACMG/AMP standards and guidelines. The diagnosis was confirmed by genetic testing and pathogenicity was established by clinical fibroblast testing and functional assays. Based on our observations, COL3A1 exon19:c.1340G > A variant may now be classified as a likely pathogenic variant for vEDS exome analysis. Further detailed study may be required to understand the exact molecular mechanism of vEDS pathogenesis in this variant.
The original contributions presented in this study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.
Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. The patients/participants provided their written informed consent to participate in this case study. Written informed consent was obtained from the individual(s) for the publication of this case study and any potentially identifiable images or data included in this article.
JM, LL, and UK conceived and designed the study. LL and UK performed the clinical evaluation. JM, AS, and SS performed and analyzed the experiments related to genetic testing and biochemistry. SG analyzed the radiology findings. JM, LL, and AS wrote the article. All authors proofread the manuscript. | true | true | true |
|
PMC9596146 | 36111549 | Andreas Enström,Robert Carlsson,Ilknur Özen,Gesine Paul | RGS5: a novel role as a hypoxia-responsive protein that suppresses chemokinetic and chemotactic migration in brain pericytes | 17-10-2022 | Hypoxia,Migration,PDGFBB,Pericytes,RGS5,S1P | ABSTRACT Adaptive biological mechanisms to hypoxia are crucial to maintain oxygen homeostasis, especially in the brain. Pericytes, cells uniquely positioned at the blood-brain interface, respond fast to hypoxia by expressing regulator of G-protein signalling 5 (RGS5), a negative regulator of G-protein-coupled receptors. RGS5 expression in pericytes is observed in pathological hypoxic environments (e.g. tumours and ischaemic stroke) and associated with perivascular depletion of pericytes and vessel leakage. However, the regulation of RGS5 expression and its functional role in pericytes are not known. We demonstrate that RGS5 acts as a hypoxia-responsive protein in human brain pericytes that is regulated independent of hypoxia inducible factor-1α (HIF-1α), rapidly stabilized under hypoxia, but degraded under normoxic conditions. We show that RGS5 expression desensitizes pericytes to signalling of platelet-derived growth factor-BB (PDGFBB) and sphingosine 1-phosphate (S1P), and blocks chemokinesis or chemotaxis induced by these factors. Our data imply a role for RGS5 in antagonizing pericyte recruitment and retention to blood vessels during hypoxia and support RGS5 as a target in counteracting vessel leakage under pathological hypoxic conditions. This article has an associated First Person interview with the first author of the paper. | RGS5: a novel role as a hypoxia-responsive protein that suppresses chemokinetic and chemotactic migration in brain pericytes
Adaptive biological mechanisms to hypoxia are crucial to maintain oxygen homeostasis, especially in the brain. Pericytes, cells uniquely positioned at the blood-brain interface, respond fast to hypoxia by expressing regulator of G-protein signalling 5 (RGS5), a negative regulator of G-protein-coupled receptors. RGS5 expression in pericytes is observed in pathological hypoxic environments (e.g. tumours and ischaemic stroke) and associated with perivascular depletion of pericytes and vessel leakage. However, the regulation of RGS5 expression and its functional role in pericytes are not known. We demonstrate that RGS5 acts as a hypoxia-responsive protein in human brain pericytes that is regulated independent of hypoxia inducible factor-1α (HIF-1α), rapidly stabilized under hypoxia, but degraded under normoxic conditions. We show that RGS5 expression desensitizes pericytes to signalling of platelet-derived growth factor-BB (PDGFBB) and sphingosine 1-phosphate (S1P), and blocks chemokinesis or chemotaxis induced by these factors. Our data imply a role for RGS5 in antagonizing pericyte recruitment and retention to blood vessels during hypoxia and support RGS5 as a target in counteracting vessel leakage under pathological hypoxic conditions. This article has an associated First Person interview with the first author of the paper.
Oxygen is fundamental for the survival and normal function of most multicellular organisms. Therefore, specific cellular mechanisms have evolved to sense and respond to hypoxia so that oxygen homeostasis is maintained within tissues and organs. The brain consumes ∼20% of the total available oxygen at rest, and, as a result, the brain vasculature has adapted to meet its high oxygen and energy demand, resulting in a 400 mile-long capillary network (Cipolla, 2009). Pericytes are perivascular cells that line the entire microvasculature of the brain and one of the key cell types that maintain blood-brain barrier (BBB) integrity and regulate the formation of new blood vessels (Armulik et al., 2010; Eilken et al., 2017). Due to their unique position at the blood-brain interface, pericytes are known to be one of the first responders to hypoxia (Duz et al., 2007; Gonul et al., 2002). During vascular development, pericytes are essential for vessel stabilization and maintenance (Daneman et al., 2010; Hellstrom et al., 1999), whereas in adulthood and under pathological conditions, they play an active role in modulating angiogenesis and maintaining the BBB integrity (Armulik et al., 2010; Kang et al., 2019; Ozerdem and Stallcup, 2003). These processes are often initiated as an adaptive response to hypoxic environments and are associated with the expression of regulator of G-protein signalling 5 (RGS5) (Mitchell et al., 2008), a protein that in the brain is exclusively expressed by vascular mural cells [pericytes and smooth muscle cells (SMCs)] (Bondjers et al., 2003; Cho et al., 2003). RGS5 belongs to the R4 subfamily of RGS proteins and is a cytoplasmic protein that acts as a negative regulator of G-protein-coupled receptors (GPCRs) (De Vries et al., 2000). Recent discoveries have also highlighted non-GPCR-related intracellular targets, where RGS5 exerts a regulatory role (Hauser et al., 2018; Wang et al., 2021). RGS5 expression in pericytes in the adult brain is usually low under physiological conditions. Interestingly, however, RGS5 increases dramatically in pathological hypoxia (e.g. in tumours or ischaemic stroke), where it is associated with pericyte detachment and migration from the capillaries into the brain parenchyma, resulting in BBB leakage (Hamzah et al., 2008; Ozen et al., 2014, 2018; Roth et al., 2019). Those observations suggest that RGS5 is a prominent protein in the process of vascular remodelling and appears to play a regulatory role in the association of pericytes to the microvasculature, although the underlying mechanisms are not known. Pericyte detachment from the microvasculature is a necessary process during vascular remodelling, allowing endothelial sprouting before subsequent initiation of pericyte migration and recruitment for stabilization of newly formed vessels (Kamouchi et al., 2012). The recruitment and retention of pericytes to the microvasculature is actively modulated by chemotactic factors secreted by endothelial cells, such as platelet derived growth factor-BB (PDGFBB) and sphingosine 1-phosphate (S1P) (Aguilera and Brekken, 2014; Jain, 2003; Karakiulakis et al., 2007; Renner et al., 2003; Wittko-Schneider et al., 2014). Previous studies on targeted deletion of PDGFBB and S1P signalling demonstrate an increase in pericyte dissociation from the vascular wall and lack of pericyte recruitment (Kono et al., 2004; Lindblom et al., 2003). However, whether the detachment or lack of recruitment of pericytes from microvessels in pathological hypoxia is related to disruption of pericyte chemotactic cues is unclear. The previous observations led us to hypothesize that the expression of RGS5 in hypoxic pericytes may play a regulatory role in PDGFBB- or S1P-induced pericyte recruitment and migration. Here, we provide a detailed analysis of the regulation and the functional role of endogenous RGS5 expression in human brain pericytes under pathological hypoxic conditions. We show that RGS5 acts as a hypoxia-responsive protein in pericytes that is regulated independent of hypoxia inducible factor-1α (HIF-1α) and stabilized under hypoxia, whereas it is rapidly degraded under normal oxygen conditions in vitro. We examine the function of RGS5 under hypoxic conditions using siRNA silencing and characterize proliferation, apoptosis, chemokinetic and chemotactic migration of brain pericytes as well as downstream mitogen-activated protein kinase (MAPK) signalling after PDGF receptor-β (PDGFRβ) and S1P receptor (S1PR) stimulation. We demonstrate that RGS5 expression under hypoxia reduces PDGFRβ phosphorylation and desensitizes pericytes to the chemotactic and chemokinetic response induced by PDGFBB, while this effect is not observed under normoxic conditions. In addition, RGS5 inhibits the chemokinetic migratory response of pericytes to S1P. These findings suggest a role for RGS5 in antagonizing pericyte retention and/or recruitment to endothelial cells during hypoxia and vascular remodelling.
Based on our previous observations of increased RGS5 expression in brain pericytes in hypoxic/ischaemic pathology in vivo (Ozen et al., 2014, 2018; Roth et al., 2019), we hypothesized that RGS5 is mainly present and has its primary function in hypoxic environments. Confirming our hypothesis, we found that endogenous RGS5 protein expression was extremely low in normoxia (Fig. 1A,B). In contrast, when we exposed pericytes to hypoxia, we detected a time-dependent induction in RGS5 protein levels. We observed a ∼5-fold increase in RGS5 as early as after 1 h, which gradually increased and reached its peak at 12 h before returning towards baseline at 24 h, confirming that RGS5 is hypoxia inducible (Fig. 1B). Similarly, hypoxia increased the protein levels of HIF-1α, a known hypoxia-inducible transcription factor, in a time-dependent manner, showing a continuous increase and peak at 6 h before returning towards baseline at 24 h (Fig. 1D). However, RGS5 mRNA expression remained relatively constant, displaying slight but even significantly lower levels of mRNA at both 1 and 24 h of hypoxia (Fig. 1C). Similarly, HIF1A mRNA remained constant except for a significant increase at 1 h of hypoxia (Fig. 1E), indicating that both hypoxia-inducible proteins are regulated primarily post-translationally.
Many of the so far discovered adaptive cellular responses to hypoxia are linked to transcriptional activation by HIFs (Majmundar et al., 2010). This has also been claimed for RGS5; however, this study was performed in human-umbilical-vein endothelial cells forced to overexpress RGS5 (Jin et al., 2009). To directly evaluate whether the transcription factor HIF-1α also regulates RGS5 expression in pericytes, we used siRNA to silence the expression of HIF1A and measured protein levels of HIF-1α and RGS5 under hypoxia. HIF1A siRNA significantly downregulated HIF-1α protein expression, confirming the efficiency of the siRNA silencing (Fig. 1F,H). Consistent with the relatively constant levels of RGS5 mRNA shown previously, siRNA knockdown targeting HIF1A did not lead to a significant change in RGS5 protein levels at any of the time points of 0, 3, 6, 12 or 24 h of hypoxia, confirming that the hypoxic induction of RGS5 is not regulated by HIF-1α-initiated transcription (Fig. 1F-H). Our findings suggest that RGS5 functions as a hypoxia-responsive protein in brain pericytes and that its protein induction following hypoxia is controlled outside of the hypoxia-regulated transcriptional network and is independent of HIF1α.
To study the role of RGS5 in human brain pericytes under hypoxia, we utilized siRNA silencing of RGS5 (siRGS5) and compared the response to pericytes treated with scrambled control siRNA (siCTRL). Successful knockdown of RGS5 was verified for both mRNA and protein expression (Fig. 2A). Scrambled siCTRL showed similar levels of RGS5 when compared to wild-type (WT) pericytes (Fig. 2A). The recruitment and retention of pericytes to newly formed vessels are necessary mechanisms for vessel maturation and rely on specific chemotactic factors secreted by endothelial cells, such as S1P and PDGFBB (Aguilera and Brekken, 2014; Jain, 2003; Raza et al., 2010). Using a microfluidic migration chamber with a stable linear concentration gradient, we examined the effect of RGS5 on S1P-induced chemotaxis of pericytes under hypoxia (Fig. 2B). We performed real-time quantitative PCR (qPCR) to analyse which of the five different S1P receptors was expressed in the pericyte cell line used in this study. We observed a clear expression of S1PR2 and S1PR3 mRNA in both normoxic and hypoxic environments that was not affected by siRGS5. However, upon hypoxic exposure for 24 h, we observed a ∼ 2-3-fold decrease in both S1PR2 and S1PR3 in pericytes. We detected a slight expression of S1PR1 but complete lack of S1PR4 and S1PR5 expression in pericytes (Fig. S1). There was no clear chemotactic effect after S1P treatment for either the forward migration index (FMI) or centre of mass distribution, nor an effect on circular distribution homogeneity of cellular trajectory endpoints, where pericytes migrated randomly independent of RGS5 expression and the S1P gradient (Fig. 2C,E-G). To verify that the lack of chemotactic migration in response to S1P was not a consequence of hypoxic exposure, we conducted the same experiment under normoxic conditions. Likewise, S1P did not induce a clear effect of chemotaxis in both the siRGS5- and siCTRL-treated pericytes (Fig. S2A, B) Next, we investigated whether RGS5 regulates the chemotactic response to PDGFBB, a strong chemoattractant factor for pericytes during vascular remodelling. Here, knockdown of RGS5 resulted in a significantly increased sensitivity of pericytes to PDGFBB, while siCTRL pericytes expressing RGS5 instead exhibited more random migration (Fig. 2D). The FMI parallel () to the gradient was higher and showed a more consistent degree of migration towards the source of the gradient in siRGS5 compared to siCTRL pericytes (Fig. 2H). Similarly, the centre of mass representing the average of all single-cell endpoints was shifted towards the PDGFBB gradient to a significantly higher degree in siRGS5 compared to siCTRL pericytes (Fig. 2I). The Rayleigh test showed a significant inhomogeneous distribution of cellular endpoints in siRGS5 compared to uniform circular distribution for siCTRL pericytes (Fig. 2J), indicating that expression of RGS5 desensitizes pericytes to the chemotactic cues of PDGFBB. Furthermore, we verified these findings by conducting the same experiment under normoxic conditions in which RGS5 is degraded. PDGFBB induced a strong chemotactic effect in both siRGS5- and siCTRL-treated pericytes, suggesting that the inhibitory effect of RGS5 on PDGFBB chemotaxis is hypoxia dependent (Fig. S2C,D). These results indicate that the knockdown of RGS5 in hypoxia rescues the PDGFBB-induced chemotaxis to ∼ 50% compared to normoxia.
Next, we investigated if RGS5 regulated the efficiency of pericyte motility at a basal level or elicited a change in S1P- or PDGFBB-induced chemokinesis. To determine the effect of RGS5 on collective pericyte migration, we performed a wound-healing scratch assay for the duration of 8 h under hypoxia. RGS5 expression did not impact pericyte motility under non-stimulated conditions (siRGS5=39.9±13%; siCTRL=39.6±17%) (Fig. 3A). However, knockdown of RGS5 in pericytes stimulated with S1P showed a significant increase in percentage of wound closure compared to siCTRL (siRGS5=60.0±11%; siCTRL=36.7±8%) (Fig. 3B). The addition of PDGFBB also resulted in a slight, but significant, increase in cellular motility in siRGS5 compared to siCTRL pericytes (siRGS5=58.8±4%; siCTRL=42.7±9%) (Fig. 3C). As the wound-healing assay is limited due to the rapid gap closure, cell-cell interactions and inconsistencies in scratch diameters, we decided to validate our findings using long-term single-cell migration analysis for 24 h under hypoxia. We calculated the mean square displacement (MSD) to evaluate the overall efficiency of migration as it considers both cellular speed and direction. Here, we observed similar MSD values between non-stimulated siRGS5 compared to siCTRL pericytes under hypoxic conditions (Fig. 3D,E). Likewise, S1P treatment did not reveal any difference in MSD after long-term single-cell migration under hypoxia, and only a slight difference after PDGFBB treatment between siRGS5 and siCTRL pericytes (Fig. 3D,E). Next, we assessed the instantaneous speed of migrating pericytes for every 30 min time interval. Despite a time-dependent trend towards a decrease in the speed of migration during 24 h of hypoxia, pericytes remained migratory throughout the time period. RGS5 did not influence the instantaneous speed of non-treated pericytes (Fig. 3F,G). However, siRGS5 compared to siCTRL pericytes migrated slightly faster after S1P and PDGFBB treatment shown from the instantaneous speed and the area under the curve (AUC) evaluation, although the effect size observed is small and may not confer any functional difference. Pericytes treated with RGS5 siRNA showed no changes in MSD values after S1P treatment and even significantly lower MSD values after PDGFBB treatment, although with an effect size of only ∼6% despite having higher instantaneous speed after S1P and PDGFBB treatment (Fig. 3E-G). Since MSD takes directionality of cellular migration into account, our results suggest that siRGS5 pericytes have more meandering cellular trajectories compared to siCTRL pericytes but increased speed of migration after stimulation with S1P or PDGFBB, in absence of a concentration gradient. This is due to the slope of the MSD curves being dependent on the persistence of the random walk of the cellular trajectory and cellular speed of migration. As the persistence of directionality diminishes, it reduces the slope of the MSD coefficients. Thus, the siRGS5 pericytes migrating at faster speed when treated with PDGFBB or S1P but exhibiting similar or even slightly lower MSD curves suggest that they also exhibit more meandering migratory trajectories. Cellular geometry is intricately linked to cellular motility. The balance between retraction and expansion of cellular protrusions alters the cellular shape of migrating cells (Chen et al., 2013). To assess if RGS5 regulated cellular shape, we measured cellular solidity as well as the area of the cell body after 24 h of hypoxia. However, no significant differences were observed between siRGS5 and siCTRL pericytes regardless of stimuli (Fig. S3). In summary, our findings demonstrate that RGS5 reduces the chemokinetic effect of S1P and PDGFBB on pericytes in hypoxia during collective and long-term single-cell migration.
Next, we investigated the effect of RGS5 on intracellular signalling targets and PDGFRβ phosphorylation. MAPK signalling via extracellular-regulated kinase (ERK1/2) and protein kinase B (AKT) phosphorylation is important for cellular proliferation and migration as ERK1/2 and AKT are downstream signalling targets that modify e.g. cell cycle S-phase entry and cytoskeletal polarity (Bonacchi et al., 2001; Kluk and Hla, 2002; Mebratu and Tesfaigzi, 2009). As MAPK signalling occurs downstream of both G-protein S1PRs and PDGFRβ, we wanted to assess the potential role of RGS5 as a regulator of their activation in brain pericytes under hypoxic conditions (Cho et al., 2003; Gunaje et al., 2011). Therefore, we treated siRGS5 or siCTRL pericytes with PDGFBB or S1P for 5 min after the cells had been exposed to 8 h of hypoxia. We observed a significant increase in phosphorylated (p)-ERK in siRGS5 compared to siCTRL pericytes upon S1P stimulation but no significant effect on p-AKT (Fig. 4A,B). However, these results display a large standard deviation and should be interpreted with caution. PDGFBB-initiated PDGFRβ phosphorylation induced a significantly higher degree of phosphorylation of both Y751 and Y1021 tyrosine residues on the PDGFRβ intracellular domain in siRGS5 compared to siCTRL pericytes (Fig. 4C). However, no significant difference was observed in p-ERK or p-AKT signalling between siRGS5 and siCTRL pericytes after PDGFBB treatment (Fig. 4C,D). Non-treated pericytes showed no significant difference in basal p-ERK nor p-AKT activation between siRGS5 and siCTRL pericytes. Furthermore, non-treated cells elicited undetectable levels of both Y1021 and Y751 of p-PDGFRβ regardless of RGS5 expression under hypoxia (Fig. 4E,F). In addition, we evaluated the signalling targets in normoxia, where we did not detect a significant difference between siRGS5 and siCTRL cells, after S1P or PDGFBB treatment or under non-treated conditions (Fig. S4A-E). Our results displayed a trend for increased p-ERK after both PDGFBB and S1P stimuli, but this was not significant compared to non-treated conditions (Fig. S4B). However, PDGFBB treatment had a clear effect on AKT and PDGFRβ phosphorylation (Fig. S4C-E). The importance of ERK phosphorylation in cell cycle entry throughout G0/G1 to facilitate S-phase progression is well documented (Mebratu and Tesfaigzi, 2009). Likewise, PDGFRβ activation is known to induce pericyte proliferation and survival (Payne et al., 2021; Renner et al., 2003). As we had shown that RGS5 interferes with p-ERK and p-PDGFRβ after S1P or PDGFBB treatment, respectively, we next investigated whether RGS5 has the potential of regulating pericyte proliferation or apoptosis.
Even though hypoxia induces cell cycle arrest in most cells, certain cell types that reside in the hypoxic niche, including endothelial cells, SMCs and pericytes, may retain their proliferative capacity during e.g. hypoxia-induced angiogenesis (Hubbi and Semenza, 2015; Namiki et al., 1995; Phillips et al., 1995; Lannér et al., 2005). We have previously observed that loss of RGS5 in vivo resulted in increased numbers of pericytes in the infarct core 24 h after ischaemic stroke (Ozen et al., 2018). We hypothesized that this effect could, at least in part, be the consequence of RGS5 regulating pericyte proliferation or apoptosis. Therefore, we analysed the different stages of the cell cycle using 5-ethynyl-2′-deoxyuridine (EdU) incorporation and nuclear staining by flow cytometry. Despite that pericytes retained proliferative activity after 8 h of hypoxia, knockdown of RGS5 did not lead to significant changes in cell cycle progression of G0/G1, S, G2 or mitosis compared to siCTRL pericytes under non-stimulated conditions (Fig. 5A). Furthermore, S1P or PDGFBB treatment resulted in similar pericyte proliferation profiles to non-stimulated conditions regardless of RGS5 expression (Fig. 5A-C). To determine whether loss of RGS5 may instead affect apoptosis in pericytes, we also evaluated the sub-G1 peak corresponding to nuclear fragmentation of apoptotic cells. We did not observe any significant difference between siRGS5 and siCTRL pericytes, with or without PDGFBB or S1P stimuli (Fig. 5A-C). Our results indicate that pericytes’ propensity to complete cellular division or undergo apoptosis after PDGFBB or S1P treatment under hypoxic conditions remained unchanged compared to that of non-stimulated cells and is not affected by RGS5 expression.
Understanding the molecular mechanisms governing cellular oxygen sensing is necessary to discern how cells respond to hypoxia and the pathophysiological implications thereof. It also allows the exploration of novel targets to modulate these processes. When hypoxia occurs in the brain, pericytes that are uniquely located at the blood-brain interface are one of the first responders and express high levels of RGS5 (Duz et al., 2007; Ganss, 2015; Gonul et al., 2002; Ozen et al., 2018; Roth et al., 2019). Here, we show that RGS5 acts as a hypoxia-responsive protein in brain pericytes. We demonstrate that endogenous RGS5 protein in brain pericytes in vitro is present and stabilized under hypoxic conditions. Previously, RGS proteins have often been studied under normoxic conditions in cells with otherwise little endogenous expression of RGS5 using overexpression models (Ganss, 2015; Jin et al., 2009). Whilst studies using vector expression of RGS5 may be a helpful tool to investigate molecular mechanisms regulated by RGS5, they introduce highly enriched protein levels and in conditions in which it may be physiologically irrelevant, such as in normoxia. To our knowledge, this is the first time that RGS5 is identified as a hypoxia responder in brain pericytes in which RGS5 protein expression is shown to be dependent on a hypoxic environment. Our results are in line with recent evidence demonstrating that RGS5 is a subject of the NO/O2-dependent Arg/N-degron pathway and undergoes rapid proteolysis in the presence of oxygen due to its N-terminal configuration. In the case of RGS5, when oxygen is available, the cysteine residue at the N-terminal is oxidized and recognized by arginyl-tRNA-protein transferase, which transfers arginine to the cysteine residue and primes RGS5 for ubiquitin ligation and degradation (Lee et al., 2005; Masson et al., 2019; Ganss, 2015). Thus, in normoxia, GPCR and other signalling mechanisms can be maintained due to the absence of RGS5 that would otherwise inhibit receptor downstream activation. It has been postulated that the conditionally short half-life of RGS5 under normoxia is a cellular mechanism to ensure maximal responsiveness to environmental changes like hypoxia (Lee et al., 2012). Thus, in a hypoxic environment, the N-terminal oxidation of RGS5 is inhibited and the degradation process stopped. Our results demonstrate that the induction of RGS5 in hypoxia is exclusively regulated post-translationally, where RGS5 stabilization is a rapid real-time response to hypoxia and independent of HIF-1α-mediated transcriptional activation. On a physiological level, it is likely that the direct post-translational stabilization of RGS5 leads to faster adaptations to hypoxia than the transcriptional responses transduced by HIF-1α in brain pericytes. As pericytes are crucial for vascular stability and remodelling during vasculogenesis and angiogenesis, we investigated the effect of RGS5 on S1P and PDGFBB signalling, known to be key angiogenic factors that regulate and coordinate pericyte recruitment, retention and proliferation. Pericytes typically express S1PR1, S1PR2 and S1PR3 (He et al., 2018; Uemura et al., 2020; Vanlandewijck et al., 2018). Previous studies support the notion of S1P signalling being important for mural cell association to blood vessels (Cartier et al., 2020; Kono et al., 2004). S1P signalling also has chemotactic responses in various cell types including macrophages and SMCs (Duru et al., 2012; Keul et al., 2011), but the contribution of S1P signalling in pericytes during vascular formation and pericyte recruitment is not clear. We observed expression of S1PR1-S1PR3 in our pericyte cell line, which can signal via the Gi or Gq GPCR pathway but can induce opposing cellular effects. For example, overexpressing S1PR1 induces enhancement of mural cell coverage, while knocking out S1PR2 yields similar results (Cartier et al., 2020). Since RGS5 interacts with Gi and Gq signalling pathways we wanted to test our hypothesis that S1P signalling in pericytes may induce chemotaxis, where RGS5 could play a regulatory role. In line with these results, previous studies have shown a link between S1P signalling and induction of SMC migration (Böhm et al., 2013; Liu et al., 2018). Studies performed under normoxic conditions on human aortic SMCs overexpressing RGS5 or in rat aortic SMCs with endogenous RGS5 expression have suggested that RGS5 attenuates S1PR signalling (Cho et al., 2003; Gunaje et al., 2011). Here, we demonstrate that RGS5 significantly inhibits downstream phosphorylation of ERK but did not have a significant effect on p-AKT after S1P treatment in brain pericytes in the context of pathological hypoxia. This suggests that RGS5 may have a regulatory role in the intrinsic GTPase activity of the G-protein-coupled S1PRs. However, these results display a large standard deviation and should be interpreted with caution, as the role of RGS5 in MAPK signalling after S1P stimulation requires further investigation. Furthermore, we demonstrate that RGS5 elicits an inhibitory function in S1P-induced chemokinesis; however, S1P did not have a chemotactic effect on pericytes. The initiation of vascular remodelling requires pericyte detachment from the vessel wall, allowing endothelial sprouting before pericytes are recruited back to the endothelial cell layer to achieve vessel maturation (Kamouchi et al., 2012). Pericytes thereby respond to the chemotactic cues of PDGFBB secreted by endothelial cells (Aguilera and Brekken, 2014). PDGFBB, once secreted, stays normally bound to the extracellular matrix because of its C-terminal retention motif (Ostman et al., 1991). Targeted deletion of the retention motif leads to partial dissociation of pericytes and cellular protrusions extending away from the vascular wall (Lindblom et al., 2003). This confirms that PDGFBB is not only necessary for pericyte recruitment but is also important for pericyte retention to the endothelium. Interestingly, our results demonstrate that RGS5 reduces the sensitivity of pericytes to PDGFBB-induced chemokinesis and chemotaxis, thereby antagonizing recruitment and retention of pericytes to the endothelium under hypoxic conditions. These results are in line with our recent studies in an in vivo model of ischaemic stroke, showing that pericytes respond early to stroke and express RGS5 before they detach from the endothelial cell layer and migrate into the brain parenchyma, causing perivascular depletion of pericytes. Under these pathological conditions, this originally adaptive process of pericyte detachment during vascular remodelling leads to BBB breakdown, increased oedema and aggravation of neuronal cell death (Ozen et al., 2014, 2018). In turn, we and others have previously shown that when RGS5 is knocked out in models of ischaemic stroke (Ozen et al., 2018; Roth et al., 2019) or tumours (Hamzah et al., 2008), pericyte coverage of the vascular wall and vessel integrity are preserved and vascular leakage is substantially reduced. Thus, our present data demonstrate that expression of RGS5 is a possible mechanism for pericyte detachment or lack of pericyte recruitment as an adaptive response to hypoxic conditions by counteracting PDGFBB-induced chemotactic cues. This is further demonstrated by the strong PDGFBB chemotactic effect in normoxia regardless of siRGS5 or siCTRL treatment due to the constant degradation of RGS5 in the presence of oxygen. Interestingly, knockdown of RGS5 in hypoxia conserves ∼ 50% of the PDGFBB-induced chemotaxis; however, in the RGS5-expressing siCTRL pericytes, this effect was lost. Even though the focus of this study is the regulation of RGS5 and its functional role in hypoxia, we also investigated some of the possible underlying mechanisms of the reduced responsiveness to PDGFBB under hypoxia. Specifically, we show that RGS5 expression is associated with inhibition of phosphorylation of the tyrosine intracellular domains Y751 and Y1021 of PDGFRβ upon PDGFBB stimulation. However, we did not detect any difference in phosphorylation of the downstream targets ERK or AKT after PDGFRβ activation, suggesting that PDGFBB-induced chemotaxis and chemokinesis are regulated via other downstream signalling trajectories or different mechanisms under hypoxia (Dubrac et al., 2018). Because PDGFRβ phosphorylation and dimerization after ligand binding is dependent on several scaffolding proteins to induce downstream signalling networks with the potential to induce pericyte migration (Dubrac et al., 2018; Aguilera and Brekken, 2014), further studies that e.g. evaluate the phospho-proteomic profile regulated by RGS5 in response to PDGFBB are needed. While RGS5 expression impairs pericyte migration and recruitment in vitro, pericyte proliferation remained unchanged, regardless of stimuli and RGS5 expression under hypoxia. Even under normal physiological conditions, the initiation of cellular division is a tightly controlled and a highly energy-consuming process that requires cellular sensing mechanisms to regulate and facilitate cell cycle progression. Our results show that pericytes instead favour a migratory phenotype over proliferation alterations under pathological hypoxia. Hypoxia promotes the highly coordinated angiogenic or vasculogenic response. It is possible that the susceptibility of RGS5 stabilization/degradation to fluctuations in oxygen levels has a role in fine tuning the process of vascular remodelling. In vascular pathology like ischaemic stroke, tissue oxygenation is severely disturbed, leading to severe hypoxia and likely fast stabilization kinetics of RGS5, where its effects remain mostly unexplored. It is conceivable that the stabilization of RGS5 under prolonged hypoxia may lead to a dysregulation in pericyte signalling during vascular remodelling and possibly contributes to perivascular pericyte loss in hypoxic/ischaemic brain pathology. In summary, we demonstrate that RGS5 is a hypoxia-responsive protein in human brain pericytes that is regulated independent of HIF-1α and allows a rapid real-time response to hypoxia able to decouple pericytes from specific extracellular signals related to pericyte recruitment and retention to the vasculature. Thus, RGS5 may constitute a target in pericytes to modulate unbalanced responses to hypoxia under pathological situations.
Primary human brain pericytes isolated from cerebral cortex tissue were purchased from Cell-systems (ACBRI 498), where they were tested for bacterial, fungal and mycoplasma sterility by an independent laboratory. The pericyte cell line has been verified by immunofluorescence assays in which >98% were positive for desmin at passage (P)3, PDGFRB at P3 and P10, NG2 at P3 and P10, CD13 at P3 and P10, and α-SMA at P10, while <2% were positive for CD31, MAP2, neurofilament neuronal marker s100A4, GFAP, GS astrocyte marker, CD11b and Iba1 for pericyte characterization. The cells were grown in complete-classic medium supplemented with 10% serum, 5 ml CultureBoost and (2 ng/ml) Bac-Off at 37°C and 5% CO2, coated with the attachment factor derived from heat-inactivated charcoal-stripped fetal bovine serum in PBS-based HEPES-buffered gelatine vehicle (Cell-systems). Cells from passage 3-6 were used for experiments.
Cells were seeded in culture plates coated with attachment factor at 40-80% confluency as specified for each experiment and left to adhere overnight (O/N). Lipid siRNA complexes were generated with lipofectamine RNAimax reagent (Life Technologies), and siRNA transfection was conducted according to the manufacturer’s protocol. Briefly, cells were washed with PBS and the culture medium was replaced with serum-reduced Opti-MEM medium (Thermo Fisher Scientific). The siRNA complexes and lipofectamine RNAimax were diluted in Opti-MEM before being added in 1:1 ratio and left to incubate at room temperature (RT) for 5 min. Concentrations of diluted siRNA-lipid complexes were scaled according to the manufacturer’s protocol and added to the cells with the final concentrations ranging from 5 to 25 pmol siRNA and 1.5 to 7.5 μl lipofectamine (depending on the number of cells seeded) in 24-well, 12-well or six-well plates. After 6 h at 37°C in the incubator, the medium was replaced with normal complete-classic medium supplemented with 10% serum, 5 ml CultureBoost and Bac-Off (2 ng/ml). The following siRNAs were used: RGS5 silencer siRNA (Ambion, 4392420), HIF-1α silencer siRNA (Ambion, am51331) and silencer negative control #1 siRNA (Ambion, 4390843). Additionally, we used non-siRNA-treated cells as WT control group when indicated.
Cells were seeded in a 24-well plate coated with attachment factor at 30% confluency and incubated O/N at 37°C and 5% CO2. Cells were transfected with RGS5 siRNA or control siRNA and left to incubate for 24 h before the start of the experiment. First, the cells were pre-incubated for 3 h under oxygen deprivation (0.5-0.7% O2) in a humidified, gas-tight hypoxia chamber (Electrotek) with a gas composition of 85% N2, 10% H2 and 5% CO2. The growth medium was replaced with serum-free Dulbecco's modified Eagle medium (DMEM/F12) supplemented with 15 mM HEPES (Thermo Fischer Scientific) and deoxygenated by pre-bubbling the medium for 15 min in N2 gas, generating an oxygen concentration of 0.5-0.7% in the medium when added to the cells. Throughout experiments, Electrotek anaerobic indicator solution was used (containing, 2% w/v C6H12O6, 9% w/v NaHCO3 and 1% w/v Methylene Blue solution in water) to monitor O2 levels below 1% (Electrotek-scientific). After 3 h, cells were stimulated with 1 µM S1P conjugated to human serum albumin as a carrier protein (Avanti Polar Lipids) or 5 ng/ml PDGFBB (R&D Systems) (Gunaje et al., 2011; Liu et al., 2018; Yu et al., 2000). Next, cells were directly transferred to the live-imaging microscopy platform cell discoverer CD7 (Zeiss microscopy) with a humidified and hypoxic atmosphere at 37°C in 5% CO2 and 0.1% O2. Cells were imaged, tracked and averaged from four pre-selected non-overlapping positions per well with 30 min frame intervals for a duration of 24 h using phase contrast and a 5× objective. Cells from three independent experiments were analysed.
Cellular trajectories were created by manually tracking cells using the MTrackJ plug-in in ImageJ (National Institutes of Health). The x, y coordinates were used for further analysis. For quantification of MSD and speed assessment, an open-source computer program (DiPer) and open program source codes were used (Gorelik and Gautreau, 2014). MSDs were computed by the software according to Eqn 1 using overlapping time intervals. Here, MSD (n) is the MSD for a specified cell for step size n and N is the total number of cell displacements per trajectory. Here, Δt refers to the time interval between adjacent points along the migration trajectory. Population averages were then generated for each time interval nΔt where MSDs were averaged over all tracked cells and plotted against the time interval.The instantaneous speed was calculated by the software for each cell that was tracked (ν=d/t) for each time frame. Additionally, the AUC was calculated from the instantaneous speed and plotted as bar graphs.
For hypoxic imaging, pericytes (10,000 cells) were seeded on fibronectin (10 μg/ml)-coated chemotaxis µ-slides (Ibidi) and left to adhere O/N. Prior to live imaging, the chemotaxis µ-slides were transferred to the hypoxic chamber, and the reservoirs were filled with deoxygenated (0.5-0.7% O2) serum-free DMEM/F12 supplemented with 15 nM HEPES. Chemo-attractants PDGFBB (50 ng/ml) or S1P (1 μM) were applied to the left reservoir after 1 h of hypoxic pre-incubation. The µ-slides were then imaged using the CD7 microscope with a humidified and hypoxic atmosphere at 37°C in 5% CO2 and 0.1% O2. The chemotactic live imaging acquisition was then conducted during 1-23 h of hypoxia. For normoxic imaging, the same experimental set-up was used, but the live imaging was performed using a Nikon eclipse Ti microscope with a humidified chamber from Okolab at 37°C in 5% CO2 with atmospheric oxygen concentrations. Time frame intervals were set to 15 min for the duration of 22 h using a 20× objective with 0.5× magnification for hypoxic imaging, and a 10× objective was used for the normoxic microscope. Subsequently, an equal number of cells from the experimental groups siRGS5 or siCTRL pericytes treated with either PDGFBB or S1P were analysed from three independent experiments. The cells from each experiment were tracked using ImageJ plugin MTrackJ throughout the time period. The x and y coordinates from the cell tracks were exported to txt format and directly imported, statistically analysed, and computed using the chemotaxis and migration tool 1.01 (Ibidi integrated BioDiagnostics) in ImageJ. The cell trajectories were all extrapolated to x, y=0 at the starting point of time 0 h.
Pericytes were seeded at 60% confluency (25,000 cells/well) on 24-well plates coated with attachment factor and left in the culture O/N. Scratches were introduced 24 h post-transfection at∼100% confluency using a 200 μl pipette tip, and the wells were washed with PBS 3× to avoid cells re-adhering. The scratch was imaged immediately afterwards at time 0 h. Subsequently, cells were transferred to the hypoxic chamber, and deoxygenated (0.5-0.7% O2) serum-free DMEM/F12 medium supplemented with 15 nM HEPES was added to the wells. Next, cells were deprived of O2 and, after 1 h, treated with either 1 μM S1P or 5 ng/ml PDGFBB. The scratches were imaged again after 8 h of hypoxic insult using 2× objective phase contrast on an Olympus CKX41 microscope. The migration was analysed with ImageJ to determine the rates of cell migration into the scratch wound-healing area by measuring the percentage of wound-healing closure between 0 and 8 h.
Proliferation evaluation was conducted using Click-iT Plus EdU Flow Cytometry Assay (Invitrogen), according to the manufacturer’s protocol. Briefly, 200,000 cells were seeded in six-well plates coated with attachment factor. Next, 24 h post-siRNA transfection, the cells were deprived of O2 (0.5-0.7% O2) for 1 h before being stimulated with 1 μM S1P or 5 ng/ml PDGFBB and left for 8 h. During the last hour, 10 μM of EdU was added to the culture medium before the cells were washed 1× with PBS and dislodged using a cell scraper. Next, cells were washed with 1 ml of PBS with bovine serum albumin (BSA) followed by another wash with PBS. Cells were then resuspended in 1 ml PBS with LIVE/DEAD Fixable Dead Cell Stain kit (1:1000, Thermo Fisher Scientific) and incubated on ice for 30 min before washing once with 1 ml of PBS. The cells were then washed with 1 ml of 1% BSA in PBS before addition of 100 μl of 4% paraformaldehyde (PFA) and incubation at RT for 15 min. Cells were then washed with 100 µl of 1% BSA in PBS followed by resuspension in 100 μl of 1× Click-iT permeabilization and wash reagent and incubated for another 15 min at RT. Then, cells were incubated with NG2 antibody (Thermo Fisher Scientific) at 1 μg/ml for 15 min at 4°C, washed with 100 μl of 1% BSA in PBS and resuspended in 33 μl of 1× Click-iT permeabilization and wash reagent. Thereafter, 166 μl of Click-iT Plus reaction cocktail was added to the cells for 30 min at RT before the cells were washed 2× with 100 μl of Click-iT permeabilization and wash reagent. Next, cells were resuspended in 0.5% fetal bovine serum in PBS, 4′,6-diamidino-2-phenylindole (DAPI; 1:10,000, Thermo Fisher Scientific) was added to stain the DNA content, and cells were incubated for 5 min at RT before analysis using a BD LSRII flow cytometer (BD Biosciences). Flow cytometry data were analysed using FlowJo software (FLOWJO LLC, version 10.7.2).
For western blotting, pericytes were seeded in 24-well plates (30,000 cells/well). The cells were lysed directly in the well using 100 μl of 1× Laemmli buffer (Bio-Rad) supplemented with 0.1 M DTT. The lysates were denatured at 95°C for 5 min and run on 15-well 4-15% SDS-PAGE gels (Bio-Rad) before being blotted onto Turbo-transfer-packs (Bio-Rad). The post-transfer nitrocellulose membranes were then blocked for 1 h in 5% milk in Tris-buffered saline with 0.1% Tween-20 (TBST). Next, the membranes were incubated with the primary antibodies (Table S1) in 5% BSA in TBST O/N. The membranes were subsequently washed 3× with TBST before goat-anti-rabbit-HRP secondary antibody was added (1:5000, Dako) in 5% milk in TBST and incubated at RT for 1 h. Membranes were then washed 3× with TBST followed by protein detection using HRP substrates Clarity or Clarity Max (Bio-Rad) to measure the chemiluminescence on a ChemiDoc (Bio-Rad). For evaluating signalling mechanisms of PDGFRβ and intracellular phosphorylation of MAPK targets, siRGS5 or siCTRL pericytes were exposed to 8 h of hypoxia in deoxygenated (0.5-0.7% O2) serum-free DMEM/F12 medium supplemented with 15 nM HEPES before being stimulated with either S1P (1 μM) or PDGFBB (5 ng/ml) for 5 min, directly lysed and analysed by western blotting. For additional information on antibody validation, see Table S3.
Pericytes were seeded in 24-well plates (30,000 cells/well) and total RNA was extracted using an RNeasy Mini kit (Qiagen) according to the manufacturer’s protocol. Nanodrop 2000c (Thermo Fisher Scientific) was used to determine the quantity and quality of the RNA with 260/280 and 260/230 ratios. The RNA was reverse transcribed using a Maxima First Strand cDNA Synthesis Kit (Thermo Fisher Scientific), total volume (20 μl). For the qPCR, 10 μl reactions were run with 2 ng of cDNA, 1× SsoAdvanced Universal SYBR Green Supermix and 250 nM of forward and reverse primers.For a list of the primers used in this study, see Table S2.
Statistical analysis was performed using GraphPad Prism software version 8.2.1 or chemotaxis and migration tool 1.01 (Ibidi integrated BioDiagnostics). All statistical tests were performed on independent experiments from the same pericyte cell line indicated as the sample size in the figure legends. The sample size (n), statistical test and P-values are indicated in the figures or figure legends. The significance (P<0.05) was calculated using one-way ANOVA with Tukey's multiple comparisons, multiple t-tests or unpaired Student's t-test.
10.1242/biolopen.059371_sup1 Click here for additional data file. | true | true | true |
PMC9596196 | 36227173 | Yanchao Hu,Yajie Fan,Chunyan Zhang,Congxia Wang | Palmitic acid inhibits vascular smooth muscle cell switch to synthetic phenotype via upregulation of miR-22 expression | 12-10-2022 | VSMC,palmitic acid,synthetic phenotype,miR-22 | Synthetic phenotype switch of vascular smooth muscle cells (VSMCs) has been shown to play key roles in vascular diseases. Mounting evidence has shown that fatty acid metabolism is highly associated with vascular diseases. However, how fatty acids regulate VSMC phenotype is poorly understood. Hence, the effects of palmitic acid (PA) on VSMC phenotype were determined in this study. The effect of the PA on VSMCs was measured by live/dead and EdU assays, as well as flow cytometry. Migration ability of VSMCs was evaluated using transwell assay. The underlying targets of miR-22 were predicted using bioinformatics online tools, and confirmed by luciferase reporter assay. The RNA and protein expression of certain gene was detected by qRT-PCR or western blot. PA inhibited VSMC switch to synthetic phenotype, as manifested by inhibiting VSMC proliferation, migration, and synthesis. PA upregulated miR-22 in VSMCs, and miR-22 mimics exerted similar effects as PA treatment, inhibiting VSMC switch to synthetic phenotype. Inhibition of miR-22 using miR-22 inhibitor blocked the impacts of PA on VSMC phenotype modulation, suggesting that PA modulated VSMC phenotype through upregulation of miR-22 expression. We found that ecotropic virus integration site 1 protein homolog (EVI1) was the target of miR-22 in regulation of VSMC phenotype. Overexpression of miR-22 or/and PA treatment attenuated the inhibition of EVI1 on switch of VSMCs. These findings suggested that PA inhibits VSMC switch to synthetic phenotype through upregulation of miR-22 thereby inhibiting EVI1, and correcting the dysregulation of miR-22/EVI1 or PA metabolism is a potential treatment to vascular diseases. | Palmitic acid inhibits vascular smooth muscle cell switch to synthetic phenotype via upregulation of miR-22 expression
Synthetic phenotype switch of vascular smooth muscle cells (VSMCs) has been shown to play key roles in vascular diseases. Mounting evidence has shown that fatty acid metabolism is highly associated with vascular diseases. However, how fatty acids regulate VSMC phenotype is poorly understood. Hence, the effects of palmitic acid (PA) on VSMC phenotype were determined in this study. The effect of the PA on VSMCs was measured by live/dead and EdU assays, as well as flow cytometry. Migration ability of VSMCs was evaluated using transwell assay. The underlying targets of miR-22 were predicted using bioinformatics online tools, and confirmed by luciferase reporter assay. The RNA and protein expression of certain gene was detected by qRT-PCR or western blot. PA inhibited VSMC switch to synthetic phenotype, as manifested by inhibiting VSMC proliferation, migration, and synthesis. PA upregulated miR-22 in VSMCs, and miR-22 mimics exerted similar effects as PA treatment, inhibiting VSMC switch to synthetic phenotype. Inhibition of miR-22 using miR-22 inhibitor blocked the impacts of PA on VSMC phenotype modulation, suggesting that PA modulated VSMC phenotype through upregulation of miR-22 expression. We found that ecotropic virus integration site 1 protein homolog (EVI1) was the target of miR-22 in regulation of VSMC phenotype. Overexpression of miR-22 or/and PA treatment attenuated the inhibition of EVI1 on switch of VSMCs. These findings suggested that PA inhibits VSMC switch to synthetic phenotype through upregulation of miR-22 thereby inhibiting EVI1, and correcting the dysregulation of miR-22/EVI1 or PA metabolism is a potential treatment to vascular diseases.
Vascular function is largely dependent on vascular smooth muscle cells (VSMCs). Different from the skeletal muscle cells or cardiomyocytes, VSMCs remain possessing remarkable phenotypic plasticity in response to multiple stimuli [1]. VSMCs switch from a contractile state to a dedifferentiated, synthetic phenotype, playing crucial roles in several vascular diseases [2–4]. The synthetic phenotype induces migration to the intima and enhances proliferation and extracellular matrix protein synthesis, thereby resulting in an impaired contractility of VSMC [5]. Therefore, exploration of the underlying mechanisms involved in VSMC phenotypic switch regulation is important in vascular diseases. There are multiple environmental stimuli have been identified as factors which lead to VSMC phenotype switch, such as growth factors, reactive oxidative species (ROS), and mechanical injury [6, 7]. Recent studies have shown that metabolites were also involved in regulation of VSMC phenotype [8]. For example, lactate, a product of glucose metabolism, was found to promote the synthetic phenotype of VSMCs, which links glucose metabolism to VSMC phenotypic switch [8]. Mounting evidence has shown that fatty acid metabolism is abnormal in vascular diseases, which plays an important role in the development of atherosclerosis and other vascular diseases [9, 10]. These advances suggest that fatty acids metabolism may play a role in regulation of VSMC phenotype. However, how fatty acids regulate VSMC phenotype is poorly understood. As the most common saturated fatty acid found in organism, palmitic acid (PA) serves as an energy source or component of partially biochemicals and cellular structures. The circulating level of PA is increased in metabolic disorders and correlated with the adverse outcomes of cardiovascular diseases [11–13]. Here, we aimed to examine the impacts and underlying mechanism of PA on VSMC phenotype. VSMC phenotype switch has been widely studied in transcriptional and epigenetic levels [14, 15]. We were very interested in the growing evidence supporting a critical role for miRNAs in regulating VSMC differentiation and phenotypic switch [16, 17]. A series of miRNAs have been reported as regulators of VSMC phenotype, including miR-21 [18], miR-22 [16], miR-23b [19], miR-100 [20], miR-124 [16], miR-133 [21], miR-143/145 [22], miR-146a [23], miR-195 [24], miR-221/222 [25] and miR-424 [26]. Here, we found that PA inhibits VSMC switch to synthetic phenotype via upregulation of miR-22. These results suggested that PA plays a role in regulation of VSMC phenotype.
Primary VSMCs were isolated from 8–10 weeks old male SD rat (weighed 170–250 g) thoracic aorta as reported previously [27]. Briefly, thoracic aortas were excised followed by phosphate buffered saline (PBS) washing for 3 times. After these, the aortic media layer was dissected, cut into pieces, and seeded onto a 6-well plate. Cells were maintained in DMEM supplemented with 10% fetal bovine serum (FBS), 1% penicillin and streptomycin at 37°C in a humidified incubator with 5% CO2 in atmosphere for 2 weeks. All animal procedures in this study were conducted in accordance with the National Institutes of Health Guidelines on the Use of Laboratory Animals, and were approved by the Xi’an Jiaotong University Second Affiliated Hospital.
The 2-color fluorescence with the LIVE/DEAD Viability/Cytotoxicity kit (Molecular Probes) was used to quantify the living and dead cells in this study as directed by the manufacturer’s protocol. Briefly, cells were harvested after treatment, washed with PBS twice, and incubated with 300 μl of live/dead solution for half an hour at 37°C in the dark room. Then, the fluorescence was read using a microplate reader (FLUOstar® Omega).
Proliferation of VSMCs was analyzed using the Click-it EdU kit (C10086, Invitrogen, USA). Briefly, cells were seeded on the slides at a density of 1.0 × 103 cells in 12-well plate each well. After treatment, cells were incubated with 50 μmol/L EdU solution at 37°C for 2 h. Then, cells were washed with cooled PBS for twice and fix 4% PFA at 4°C for 15 min. Following this, 100 μl Apollo reaction cocktail was added into cells followed by nucleus staining with Hoechst 33342 according to the manufacturer’s protocol. The fluorescence signal was then analyzed under a fluorescence microscope. EdU incorporation (%) = EdU positive cells/(EdU-positive cells + Hoechst-positive cells) ×100%.
The apoptosis of VSMCs was detected using an Annexin V-FITC apoptosis detection kit (C1062, Beyotime, China). Briefly, cells were collected after treatment, washed with cooled PBS twice, resuspended with 1 mL AnnexinV-FITC, and maintained for 10 min at room temperature according to the kit’s protocol. Following this, cells were subjected to flow cytometry analysis.
VSMCs were seeded in the upper chamber of transwell (12 μm) and placed in a 24-well plate at a density of 1.0 × 105 cells/well in 200 μl DMEM contained with 0.5% FBS. The lower chamber was filled up with 600 μl DMEM contained with 10% FBS. After incubation with for 24 h, medium was discarded and the lower chamber membrane was fixed with methanol at room temperature for 15 min. Subsequently, cells were stained with 0.1 crystal violet-methanol solution for 15 min at room temperature. Finally, the migrated cells were pictured and calculated under a light microscope.
RNAiso Plus reagent (Code No.: 9108, Takara) was used for the RNA isolation as the manufacturer recommended. The cDNA was synthesized using the isolated RNA (500 ng/sample) and amplification of certain genes was performed using a SYBR Green PCR kit (Takara) in a CFX200 (Bio-Rad) with the cycles of 95°C for 10 min and 40 cycles of 95°C for 5 s, 58°C for 30 s, and 72°C for 10 s. The mRNA level of each gene was normalized to housekeeping gene, namely, GAPDH or U6. The primer sequences are listed in Supplementary Table 1.
MiR-22 mimics (5′-AAGCUGCCAGUUGAAGAACUGU-3′), miR-23b mimics (5′-AUCACAUUGCCAGGGAUUACCAC-3′), miR-125b mimics (5′-UCCCUGAGACCCUAACUUGUGA-3′), negative control mimics (NC mimics, #miR1N0000001-1-10), miR-22 inhibitors (5′-ACAGUUCUUCAACUGGCAGCUU-3′), and NC inhibitors (#miR2N0000001-1-10) were synthesized by RIBOBIO Co., Ltd Chin (Guangzhou, China). Empty vector (pcDNA3.1) and EVI1 overexpression plasmid (pcDNA3.1-EVI1 OE) were purchased from GeneChem (Shanghai, China). miRNA mimics (100 nmol/L), inhibitors (200 nmol/L), or NC (5′-UUCUCCGAACGUGUCACGUTT-3′) (100 nmol/L) were transfected using Lipofectamine™ 3000 (Invitrogen) according to manufacturer’s instruction. After 60 h post-transfection, the transfected cells were harvested and utilized for further analyses.
Wt and Mt ecotropic virus integration site 1 protein homolog (EVI1) 3′UTR sequence was acquired using PCRmethod, and then cloned into SpeI and HindIII sites of pMir-Report Luciferase vector (Applied Biosystems). The resulting construct was transfected (5ng) into 293T cells with 20 nM control mimics or miR-22 mimics using Lipofectamine-2000 (Invitrogen). After 24 h post-transfection, luciferase activity of cells was assessed using a Luciferase Assay System (Promega).
For immunoblotting, proteins were isolated from cells using RIPA buffer. Total protein extracts (15–50 μg) were separated using sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride membrane. Membranes were then probed with anti-bax (1:2000; #ab32503; Abcam), bcl-2 (1:2000; #ab196495; Abcam), cleaved-caspase-3/caspase-3 (1:2000; #ab184787; Abcam), SM22α (1:2000; #ab14106; Abcam), calponin (1:500; #ab227661; Abcam), SMMHC (1:2000; #ab125884; Abcam), vimentin (1:2000; #ab92547; Abcam), collagen I (1:1000; #ab270993; Abcam), osteopontin (OPN; 1:1000; #ab63856; Abcam), LAMC1 (1:1000; #ab233389; Abcam), EVI1 (1:1000; #SAB2100723; Sigma), AKT3 (1:2000; #ab152157; Abcam), TP53INP1 (1:2000; #ab202026; Abcam), and β-actin (1:2000; #ab8226; Abcam) at room temperature for 1.5 h. Then, membranes were immersed with the HRP-conjugated secondary antibody at room temperature for 1 h. Following this, the BM chemiluminescence blotting system (Thermo Scientific) was used for detection and protein bands were quantified using Image J software (NIH, USA).
All data are presented as mean ± standard deviation, and comparisons were performed using one-way ANOVA or two-way ANOVA followed by an unpaired t-test, as appropriate. P < 0.05 was considered statistically significant.
The datasets used and/or analyzed during the present study are available from the corresponding author upon reasonable request.
VSMCs were treated with PA (0, 100, 200 or 400 μM) for 3 d. Live/Dead assay suggested that PA treatment decreased cell viability, and increased cell death in a dose-dependent manner in VSMCs (Figure 1A). Moreover, EdU assay suggested that PA treatment significantly decreased the EdU incorporation of VSMCs in a dose-dependent manner (Figure 1B and 1C). Further analysis indicated that PA treatment could significantly increase VSMCs apoptosis (Figure 1D and 1E). Western blot analysis presented that PA treatment markedly increase the Bax and cleaved-caspase-3 expression but decreased Bcl-2 expression (Figure 1F). These findings suggested that PA may inhibit the VSMC switch to synthetic phenotype. In addition, PA treatment (200 μM) suppressed the migration of VSMCs as detected by transwell assay (Figure 1G). Furthermore, PA treatment (200 μM) for 3 d increased protein levels for markers of the contractile phenotype, including α-SMA, calponin, and SMMHC, and decreased protein levels of the synthetic phenotype, including vimentin, collagen I, and osteopontin (OPN) (Figure 1H). These results reinforced the notion that PA inhibits the VSMC switch to synthetic phenotype.
To test whether miRNA is involved in regulation of VSMC phenotype switch induced by PA, the reported miRNAs which are involved in alteration of VSMC phenotype switch were screened in PA-treated VSMCs. As shown in Figure 2A, 15 miRNAs were detected, and among these miRNAs, 3 miRNAs were increased and 1 miRNA were decreased in PA-treated VSMCs compared with that in untreated VSMCs. Following this, the top 3 increased miRNAs were overexpressed in VSMCs via using transfecting with their specific miRNA mimics, respectively (Figure 2B). Following this, the expression of synthetic and contractile markers was detected in VSMCs. As shown in Figure 2C, miR-22, miR-23b, and miR-125b mimics all increased the mRNA levels for SM22α, calponin, and SMMHC, and decreased mRNA levels of vimentin, collagen I, and OPN, suggesting that PA may inhibit the VSMC switch to synthetic phenotype via upregulation of these miRNAs. Specifically, miR-22 presented the most significant effect among these miRNAs. Thus, we had chosen miR-22 for the following investigation. Moreover, transwell assay suggested that increased expression of miR-22 mimics obviously inhibited the VSMCs migration (Figure 2D). Overexpression of miR-22 mimics also inhibited the proliferation of VSMCs (Figure 2E and 2F). In addition, flow cytometry suggested that overexpression of miR-22 mimic increased the apoptosis of VSMCs (Figure 2G and 2H). Correspondingly, the western blot analysis showed that miR-22 mimic increased the expression of Bax and clveaed-caspase-3, but decrease Bcl-2 expression (Figure 2I). These results suggested that PA may inhibit the VSMC switch to synthetic phenotype via upregulation of miR-22.
To test whether miR-22 is involved in the PA’s effects on VSMC phenotype switch, miR-22 inhibitor was used to inhibit the PA-upregulated miR-22. As shown in Figure 3A, miR-22 inhibitor decreased the miR-22 levels in VSMCs. As a result, PA treatment (200 μM) significantly inhibited the cell viability but increased apoptosis in VSMCs, while overexpression of miR-22 inhibitor attenuated the impacts of PA on the proliferation and apoptosis of VSMCs (Figure 3B–3E). Correspondingly, the western blot analysis presented that overexpression of miR-22 inhibitor attenuated the effect of PA in increasing bax and cleaved-caspase-3 expression, and decreasing bcl-2 expression (Figure 3F). Transwell analysis showed that overexpression of PA treatment significantly decreased the migration of VSMCs, but miR-22 inhibitor obviously aborted this enhancement (Figure 3G). In addition, the western blot analyses showed that PA treatment obviously accumulated the expression of SM22α, calponin, and SMMHC, but decreased the expression of vimentin, collagen I, and OPN; while overexpression of miR-22 inhibitor attenuated the effect of PA in VSMCs (Figure 3H). These results reinforced the notion that PA inhibits the VSMC switch to synthetic phenotype.
The potential target genes of miR-22 were predicted by miRDB, ENCOR1, and TargetScan. There were 50 candidates (Figure 4A), and 10 of them are associate with cell proliferation, migration, or apoptosis (Figure 4B). Among these candidates, overexpression of miR-22 reduced EVI1 mRNA levels significantly in VSMCs (Figure 4B). Western blot array showed the miR-22 mimic could significantly decrease the EVI1 expression, but miR-22 inhibitor largely enhanced the EVI1 expression (Figure 4C and 4D), indicating EVI1 acted as a candidate target of miR-22. This result was further confirmed by dual-luciferase reporter assay and presented that miR-22 reduced luciferase activity for EVI1 wild-type 3′UTR constructs but had no effect on the mutated binding site (Figure 4E and 4F). In addition, PA treatment abolished the EVI1 protein expression, while miR-22 inhibitor attenuated these downregulation, thereby upregulating EVI1 expression (Figure 4G). These results suggested that EVI1 is a target of miR-22.
To test whether EVI1 contributes to the effects of PA on VSMC phenotype switch, EVI1 and miR-22 were overexpressed in VMSCs followed by PA treatment. The qRT-PCR demonstrated that EVI1 overexpression did not change the levels of miR-22 in VSMCs, but PA treatment could enhance the upregulation of miR-22 (Figure 5A). Overexpression of miR-22 mimic significantly suppressed the EVI1 expression and PA treatment further enhanced this inhibition on the expression of EVI1 (Figure 5A). Moreover, PA treatment promoted cell proliferation as detected by EdU staining in VSMCs with EVI1 overexpression, while miR-22 mimic expression aborted this upregulation and PA treatment markedly enhanced this inhibitive effect mediated by miR-22 (Figure 5B). Flow cytometry analysis presented that overexpression of EVI1 had no obvious effect on the apoptosis of VSMCs, but miR-22 and PA treatment could enhance the apoptosis of EVI1 (Figure 5C). Similarly, EVI1 had no obviously effect on the expression of bax, bcl-2, caspase-3, and cleaved caspase-3, while miR-22 mimic and PA treatment could significantly promote the bax and cleaved-caspase-3 but decreased bcl-2 expression (Figure 5D). Transwell assays indicated that EVI1 significantly increased the migration of VSMCs, while miR-22 and PA treatment obviously attenuated this promotion to suppress the migration of VSMCs (Figure 5E and 5F). In addition, EVI1 also inhibited the contractile markers of SM22α, calponin, and SMMHC and promoted the synthetic markers of vimentin, collagen I, and OPN, while miR-22 and PA treatment attenuated these changes (Figure 5G). These results reinforced the notion that PA inhibits the VSMC switch to synthetic phenotype through regulation of miR-22/EVI1 axis.
Mounting evidence has shown that disorder in fatty acid metabolism plays a casual role in the development of atherosclerosis and other vascular diseases [9, 10]. However, how fatty acid regulates VSMC phenotype switch has not been studied. Here, we found that PA, the most common saturated fatty acid in circulation, inhibited VSMC switch to synthetic phenotype, as manifested by inhibiting VSMC proliferation, migration, and synthesis. Mechanistically, PA inhibits VSMC switch to synthetic phenotype through upregulation of miR-22 by targeting EVI1. These findings suggested that PA plays a role in the regulation of VSMC phenotype, which may contribute to vascular health and diseases. Several studies have shown that saturated fatty acids increase the risk of cardiovascular diseases [12, 28]. According to the previous study, the saturated fatty acids were usually regarded as a singular fatty acid group and they might have the same effects during the metabolism [29, 30]. However, some investigations focused on different biomarkers of risk of cardiovascular diseases found that not all SFAs exert the same effect, namely, studies do not seem to serve as a single role of PA in the development of cardiovascular diseases [31]. PA slightly elevated the LDL- and HDL-cholesterol, which is a significant predictor for cardiovascular disease [32, 33]. Although the role of PA in cardiovascular diseases needs to be further examined, these advances suggest that PA’s effects in cardiovascular health and disease cannot be easily identified as detrimental or beneficial. Here, we found that PA inhibited VSMC switch to synthetic phenotype, as manifested by inhibiting VSMC proliferation, migration, and synthesis, suggesting that PA may exert beneficial effects on vascular health and diseases, which should be identified by further studies. Recent studies support a critical role of miRNAs in regulating VSMC differentiation and phenotype switch, and miR-22 is one of the miRNAs which inhibits VSMC switch to synthetic phenotype [16, 17]. miR-22 is previously demonstrated as a tumor suppressor, but later has been concerned as a prohypertrophic miRNA [34, 35]. A recent study documented that miR-22 playing key role in the regulation role in VSMC biological activity [36]. In addition, it has also been reported that miR-22 involved in VSMC phenotypic modulation, which induces VSMC contractile gene expression, but inhibits VSMC proliferation and migration [17]. These findings indicated that miR-22 serves a key role in regulation of cardiovascular function. Here, we show that PA increased miR-22 expression in VSMCs, and inhibition of miR-22 abolished the PA’s effects on modulation of VSMC phenotype. It has been reported that transforming growth factor-β1 (TGF-β1) transcriptionally modulates miR-22 expression in VSMCs via a P53-dependent mechanism [17]. Whether PA regulates miR-22 expression through TGF-β1 needed further investigation. Indeed, there is evidence that PA treatment increases TGF-β1 in other cells [37]. These findings suggested that PA modulate VSMC phenotype via upregulating miR-22, which serves a crucial role vascular function regulation. Previous studies demonstrated that EVI1 functions as a transcriptional regulator to modulate several biological processes, including hematopoiesis, apoptosis, development, differentiation and proliferation [38, 39]. Here, we have found that EVI1 serves as a target gene of miR-22 to modulate VSMC phenotype switch. Further analysis showed that EVI1 transcriptionally inhibits VSMC-specific genes to modulate the VSMC phenotype switch, including SMαA, SM22α, SRF, and Myocd [17]. In addition, inhibiting EVI1 abolished the effects of miR-22 and PA in modulation of VSMC phenotype. These findings suggested that miR-22/EVI1 signaling axis plays a key role in VSMC phenotypic switch and correcting the dysregulation of miR-22/EVI1 or PA could be a potential treatment to vascular diseases.
Taken together, we found that PA inhibits VSMC switch to synthetic phenotype through upregulation of miR-22 expression. In addition, miR22 inhibits VSMC switch to synthetic phenotype by targeting EVI1. These findings suggested that PA plays a role in regulation of VSMC phenotype, which may contribute to maintenance of vascular health and prevention of vascular diseases. | true | true | true |
PMC9596209 | 36214767 | Fengjie Hao,Nan Wang,Honglian Gui,Yifan Zhang,Zhiyuan Wu,Junqing Wang | Pseudogene UBE2MP1 derived transcript enhances in vitro cell proliferation and apoptosis resistance of hepatocellular carcinoma cells through miR-145-5p/RGS3 axis | 07-10-2022 | pseudogene,UBE2MP1,miR-145-5p/RGS3 axis,hepatocellular carcinoma,cell growth | Pseudogenes are barely transcribed at normal, while the anomalous transcripts of them are mostly regarded as long non-coding RNAs (lncRNAs), which play potential functions in human tumorigenicity and development. The exact effects of pseudogene-derived transcripts on hepatocellular carcinoma (HCC) are ambiguous. According to our previous research and constructed database on the HCC-related lncRNAs, we noticed that UBE2MP1 was transcriptionally activated in HCC as a pseudogene from the ubiquitin-conjugating enzyme member UBE2M. In this study, we validated the high expression of the UBE2MP1 transcript in HCC and its adverse correlation with dismal outcomes for the patients. UBE2MP1 depletion at the transcript level significantly impaired cell proliferation and apoptosis resistance in HCC cell lines. Notably, we discovered that the UBE2MP1 transcript shared a specific sequence, binding to the miR-145-5p seed region with a typical ceRNA effect. Simultaneously, we verified an axis of miR-145-5p/RGS3 in HCC cells, which promoted cell proliferation and apoptosis resistance with significance. And modulation of UE2MP1 could remarkably affect RGS3 expression and consequentially influence HCC cell growth in vitro. And combined with the rescue experiment modulating either miR-145-5p or RGS3 furtherly indicated UBE2MP1 as an upstream regulator of the axis in promoting HCC cell growth and maintenance. Thus, our findings provide new strategies for HCC prevention and individual treatment. | Pseudogene UBE2MP1 derived transcript enhances in vitro cell proliferation and apoptosis resistance of hepatocellular carcinoma cells through miR-145-5p/RGS3 axis
Pseudogenes are barely transcribed at normal, while the anomalous transcripts of them are mostly regarded as long non-coding RNAs (lncRNAs), which play potential functions in human tumorigenicity and development. The exact effects of pseudogene-derived transcripts on hepatocellular carcinoma (HCC) are ambiguous. According to our previous research and constructed database on the HCC-related lncRNAs, we noticed that UBE2MP1 was transcriptionally activated in HCC as a pseudogene from the ubiquitin-conjugating enzyme member UBE2M. In this study, we validated the high expression of the UBE2MP1 transcript in HCC and its adverse correlation with dismal outcomes for the patients. UBE2MP1 depletion at the transcript level significantly impaired cell proliferation and apoptosis resistance in HCC cell lines. Notably, we discovered that the UBE2MP1 transcript shared a specific sequence, binding to the miR-145-5p seed region with a typical ceRNA effect. Simultaneously, we verified an axis of miR-145-5p/RGS3 in HCC cells, which promoted cell proliferation and apoptosis resistance with significance. And modulation of UE2MP1 could remarkably affect RGS3 expression and consequentially influence HCC cell growth in vitro. And combined with the rescue experiment modulating either miR-145-5p or RGS3 furtherly indicated UBE2MP1 as an upstream regulator of the axis in promoting HCC cell growth and maintenance. Thus, our findings provide new strategies for HCC prevention and individual treatment.
Hepatocellular carcinoma (HCC) results in severe tumor-related mortality and creates an urgent need for improving prognosis and long-term survival time for the patients [1, 2]. Targeted therapies are widely used in the clinical treatment of HCC, dependently or combined with other systematic therapies. However, the overall survival of the patients has not been accomplished to a satisfactory clinical endpoint [3]. It is important to develop more innovative therapeutic targets for the enrichment of practical HCC preventional and treating strategies. As acknowledged, a large number of transcripts exist, which are categorized into different RNA species, generated from the human genome, and the majority of these transcriptional products lack the protein-coding capacity, called the non-coding RNAs (ncRNAs). Two major classes have been identified, the small ncRNAs with the composition of fewer than 200 nucleotides, represented by the miRNAs; and the long non-coding RNAs (lncRNAs), composed with over 200 nucleotides, covering the complex mechanisms of either transcriptional or post-transcriptional events [4, 5]. Due to the large number of lncRNAs and their localization in different intracellular compartments, these kinds of molecules influence various pathological or physiological functions in the cell by interacting with DNAs, RNAs, or proteins, and are involved in gene transcription, mRNA translation, protein modification, and the formation of RNA-protein or protein-protein complexes [6, 7]. Pseudogenes are special sequences of the DNA, generated through acquired mutation and duplication of their parental genes [8, 9]. Commonly, pseudogenes are barely transcribed because of the structural defects on the promoter, premature stop codon, or the occurrence of frameshift mutation [10]. Mostly, the pseudogene-derived transcripts are considered the lncRNAs with a length of over 200 nucleotides. Taking advantage of the acknowledged mechanism by which lncRNAs exert their complex functions, we can still make sense of the pseudogene-derived transcripts functions, for example, the miRNA decoy effect, or competitive endogenous RNA (ceRNA) effect. On the other hand, the ceRNA effect is hypothesized based on the competitive interaction with the high affinity between miRNAs and the specific sequences in the lncRNAs. And this effect can abrogate the degradation of the targeted mRNAs induced by the specific miRNAs [11, 12]. A series of pseudogene-derived transcripts have been discovered to exert the ceRNA effect in human cancers and provide us with a new focus on the breakthroughs in the mechanism of HCC progress. In this study, the pseudogene of Ubiquitin-conjugating enzyme E2M (UBE2M), named UBE2M pseudogene 1 (UBE2MP1) was noticed highly transcribed in HCC with significance. Our preliminary results from the in vitro experiments demonstrated a significant correlation between UBE2MP1 and the dismal features of the patients’ clinicopathological information. And depletion of the UBE2MP1 transcript in two HCC cell lines induced remarkable defection of cell proliferation and led to a high rate of cell apoptosis. We also discovered an axis of miR-145-5p/RGS3 (Regulator of G-protein signaling 3), which promotes HCC cell growth and maintenance. And intriguingly, we explored and discovered a specific sequence in the UBE2MP1 transcript, which shares a high affinity to bind the seed region of miR-145-5p in a molecular sponge way. The modulation of either miR-145-5p or RGS3 for rescue experiments could significantly recover the phenotypes of tumor cell growth and maintenance induced by UBE2MP1. We concluded and illustrated the tumor-promoting mechanism of UBE2MP1, independent of its parental genes, by modulating the miR-145-5p/RGS3 axis. And we prompt that these findings may provide innovative and hopeful targets for HCC control.
Three typical HCC cell lines (Huh7, HepG2, and Hep3B) were recruited, and the normal human hepatic cell LO2 was used as control (Shanghai Institutes for Biological Sciences, Chinese Academy of Science, Shanghai, China)). All the cell lines were cultured by RPMI 1640, supplemented with 10% heat-inactivated fetal bovine serum (FBS), incubated at 37° C environment temperature, with 100 ug/ml streptomycin, and 100U/ml Penicillin in a humidified cell, with an atmosphere of 5% CO2. Specifically, for the transfected cells, a medium mixed with G418 (Santa Cruz Biotechnology, Inc; 400 μg/ml) was used for selection.
Nighty-three paired specimens including the tumor tissue and the adjacent non-cancerous liver tissues were collected from the patients diagnosed and conducted radical resection with no preoperative treatment, during 2016–2019, at the Department of General Surgery, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine. The corresponding clinicopathologic parameters of the patients were obtained including gender, age, tumor size, number of lesions, grades et al. The informed consent was obtained, and the study was approved by the Ethics Committee of Ruijin Hospital, Shanghai Jiao Tong University School of Medicine.
The gene expression data for 110 normal liver samples from the GTEx and TCGA along with clinical information for 369 liver tumors and 50 normal samples from the UCSC Xena database were intensively explored by using the random walk-based multi-graphic (RWMG) model algorithm developed by our team [13]. In brief, the RWMG model was developed from the biophysical interaction networks and the co-expression profiles within a single analytical framework, by which it integrates sophisticated biological connections among lncRNA targets, including the transcription factors (TF), microRNAs (miRNAs), and the alternative splice factors. This model presents flexible and scalable characteristics in ranking a subset of lncRNAs based on the literature survey. The RWMG model provides more accurate results than the previous network-based algorithms defined as the ‘shortest path’ or the conventional random walk algorithms and can avoid the ‘noise’ from the dimensional heterogenicity of the data. And the related and comprehensive information of the involved pseudogene-derived lncRNAs was described in the mentioned LCLE online. The starBase datasets (https://starbase.sysu.edu.cn/) and the dreamBase (https://rna.sysu.edu.cn/dreamBase/) datasets were used for providing supplementary information on the expression and relationship of the candidate genes in this study.
For RNA isolation, from either tissues or cells, conducted according to the instruction of the TRIzol reagent (Invitrogen, MA, USA). The first-strand cDNA was synthesized via High-Capacity cDNA Reverse Transcription Kit (ABI, NY, USA). All the primers were synthesized (Jike Biotech Company, Shanghai, China) (Supplementary Table 1). Real-time quantitative polymerase chain reaction (RT-qPCR) was operated following the TaqMan Gene Expression Assays protocol (ABI, NY, USA). The relative quantification of RNA in cell lines was normalized using GAPDH by the 2−ΔCT method. And, the relative quantification of miR-145-5p in tissue specimens and cell lines was measured by using the mirVANATM miRNA Isolation Kit (ABI, MA USA). The PCR program was set as follows: 95° C for 10 min, followed by 35 cycles of 95° C for 15s, 60° C for the 30s, and 72° C for 45s. Antibody against RGS3 was prepared (Abcam, MA, USA). The immunohistochemistry assay complied with our previously described methods [14]. The protein expression levels detected by IHC were assigned to two experienced pathologists independently for blind examination and were separated into two groups by staining intensity grade: no to low staining (0–1+) and moderate to high staining (2+–3+).
The lentiviral vectors pLKO.1 (Addgene, Cambridge, MA, USA) containing shRNA were transfected into cultured HepG2 and Hep3B cells at exponential phase (JIKE Biochemistry, Shanghai, China) for suppressing the expression of UBE2MP1 transcript. Meanwhile, the control vectors were set up. The transfected cells were selected by using a medium mixed with G418 (Santa Cruz Biotechnology, Inc; 400 μg/ml). The mimic was used to transfect HCC cells for ectopically introducing miR-145-5p (HepG2/miR-145-5p; Hep3B/miR-145-5p), and the negative controls (HepG2/NigmiR; Hep3B/NigmiR) were set. And the rescue experiment by knock-out miR-145-5p was set up by using the siRNA method by using the protocol and siRNA vector tools designed by Jike Biotech Company, (Shanghai, China), and the validation was implemented by the RT-qPCR assay. The lentiviral vector pLV (Addgene, Cambridge, MA, USA) was applied for ectopically re-expressing either RGS3 (pLV- RGS3) for the rescue experiments, and the pLV-Null was set as control.
The treated HCC cells (1x106) were cultured in 96-well microtiter plates triplicated and incubated at an atmosphere of 5% CO2 and 37° C for 5 days. Microplate computer software (Bio-Rad Laboratories, Inc., Hercules, CA, USA) was applied for measuring the OD following the Cell Counting Kit-8 (CCK-8) assay kit protocol (Dojindo, Tokyo, Japan). Then, we plotted the cell proliferation curves. Meanwhile, the cells were treated with ethanol fixation, followed by RNase A treatment and propidium iodide staining. Flow cytometry detection was carried out using FACSCalibur (Becton-Dickinson, Franklin Lakes, NJ, USA) for quantifying cell populations at the G0/G1, S, and G2/M phases, and ModFit software e (Becton-Dickinson) was used. The debris and fixation artifacts of the cells were excluded.
Cell apoptosis rate was calculated by using PE-Annexin V Apoptosis Detection Kit I (BD Pharmingen, USA) following the instructions. Transfected cells were resuspended in the concentration of 1×106 cells/ml by the 1×Binding Buffer. 5μl of FITC and 5μl of PI were added into 100μl of the cell suspension, followed by a 15 minutes incubation in darkness, added with 400μl×Binding Buffer. The apoptosis rate was calculated through flow cytometry (Becton-Dickinson). Both Annexin V-FITC-positive and PI-negative cells were considered apoptosis cells.
By using the online tools of microcosm (http://mirecords.biolead.org) and dreamBase (http://rna.sysu.edu.cn/dreamBase/), miR-145-5p was respectively predicted directly binding to the sequences of either the 3’-untranslated region (3’-UTR) of RGS3 mRNA or the UBE2MP1 transcript. Both of the sequences above were intercepted for the 202 bp sections along with the corresponding mutative ones for further detection by the dual-luciferase reporter assay. (Supplementary Table 2). The sequences were respectively cloned into the pMIR-Report luciferase vector, which contains firefly luciferase, and the pRL-TK vector luciferase was set as control (Promega, Madison, WI, USA). These two sets of vectors were co-transfected into both HepG2 and Hep3B cells transfected miR-145a-5p mimics or the control ones. The luciferase activity was measured via the Dual-Glo Luciferase assay system (Promega) 48 hours after the transfection.
Statistical analysis was carried out by using SPSS 20.0. P-values were calculated using an unpaired Student's t-test and Fisher's exact test and the one-way ANOVA as the statistical methods. Differences were considered statistically significant at P-values < 0.05.
The exploration and analysis of the datasets from the dreamBase databases indicated a universal transcription of UBE2MP1 in either pan-cancers or HCC (Figure 1A). On basis of a random walk-based multi-graphic (RWMG) model algorithm we developed, a series of pseudogenes generating lncRNA-like transcripts in HCC were screened out from the analysis tools of the Liver Cancer lncRNA Explore (LCLE) online tool (https://datasciences123.shinyapps.io/LCLE/) [13], and UBE2MP1 and the volcano plot that presents the expression profile of the pseudogene transcript indicated a similar result (Figure 1B). As we observed, merely no detectable UBE2MP1 transcript was found in the normal organs and tissues, including the liver. Whereas, the up-regulation of the UBE2MP1 transcript was detected in different tumor tissues. Especially for HCC, the remarkable elevation of the UBE2MP1 transcript was observed. The UBE2MP1 transcript expression in three HCC cell lines (Huh7, HepG2, and Hep3B) was detected by RT-qPCR assay. The level of UBE2MP1 transcript in three HCC cell lines is significantly higher than in the control LO2 cells (Figure 1C). For the real patients’ specimens from our medical center, a high level of UBE2MP1 transcript was detected in HCC tumor samples, and the adjacent non-cancerous liver tissues presented nearly no sign of UBE2MP1 transcription. As Figure 1D, 1E show, the UBE2MP1 transcript is detectable in all of the tumor specimens, and 84.95% (79/93) of the HCC specimens presented a high level of UBE2MP1 transcript expression, and the rest 15.05% (14/93) presented a relatively lower level UBE2MP1 expression. On the contrary, only 5.38% (5/93) of the adjacent non-cancerous liver tissues presented detectable but low expression of UBE2MP1. Simultaneously, we also validated the high expression of the parental gene UBE2M by RT-qPCR assay and the IHC assay in the real patient specimens and cell lines, which suggests a co-expression inclination between these two genes (Supplementary Figure 1).
The correlation between the transcription of UBE2MP1 in HCC and the clinicopathologic features of the 93 real HCC patients was analyzed statistically by Fisher's exact test and the one-way ANOVA. There is no significant correlation between UBE2MP1 and the patient’s age, gender, and virus control status. Whereas, UBE2MP1 transcription shows an adverse relationship with the tumor size (P<0.05), serum Alpha-fetoprotein (AFP) quantity (P<0.05), more advanced TNM stages (P<0.05), tumor microsatellite formation (P<0.05), invasion of venous (P<0.05), and liver cirrhosis stages (P<0.05) (Table 1). The findings here strongly suggest that the UBE2MP1 transcript plays a positive role in HCC development.
The RT-qPCR assay was conducted to validate the effect on UBE2MP1 depletion, in both HepG2 and Hep3B cells, and the expression of the parental gene was not affected (Figure 2A). The cell proliferation was significantly suppressed in both the two cell lines by abrogating the UBE2MP1 transcript (unpaired Student's t-test; *P<0.05; **P<0.01) (Figure 2B, 2C). The flow cytometric analysis showed a significant cell cycle arrest at the G0/G1 phases in the HCC cells when UBE2MP1 was depleted (Figure 2D, 2E). The percentage of the HepG2 and Hep3B cells in the G0/G1 phase increased respectively from 50.2% to 64.8% (P<0.01) and from 49.9% to 63.0% (P<0.01) (Student's t-test). Meanwhile, the obvious percentage decrease was observed in the S phase (HepG2: from 25.0% to 18.5%, P<0.05; Hep3B: from 25.7% to 17.5%, P<0.01) and the G2/M phase (HepG2: from 24.8% to 16.7%, P<0.01; Hep3B: from 24.4% to 19.5%, P<0.05).
The flow cytometric analysis was applied for calculating the statuses of cell apoptosis by an unpaired Student's t-test. The cell apoptosis rates in both HepG2 and Hep3B cells were significantly increased (HepG2: from 13.01% to 27.59%, P<0.01; Hep3B: from 13.27% to 25.71%, P<0.01) following the depletion of UBE2MP. This finding suggested that the transcription of UBE2MP1 in HCC could enhance cell maintenance by inhibiting cell apoptosis (Figure 3A–3D).
We explored the UBE2MP1 transcript and found that a sequence from 490 bp to 507 bp to the 3’ end of the UBE2MP1 transcript is probably a specific binding site that matches the seed region of miR-145-5p, which prompts a ceRNA effect of UBE2MP1 on miR-145a-5p (The minimum free energy, Mfe: -21.6 kcal/mol) (Figure 4A). The exploration of TCGA datasets also indicated a significant decline of miR-145-5p in HCC and was associated with poor clinical survival outcomes (Figure 4B, 4C). And according to the RT-qPCR assay, miR-145-5p was expressed at a relatively lower level in the HCC cell lines compared with the control LO2 cells (Figure 4D). Moreover, the expression of miR-145-5p was consequentially increased after the depletion of UBE2MP1 in the two HCC cell lines (Figure 4E). Based on this, we constructed the mutated binding site on UBE2MP1, and we carried out the dual-luciferase reporter assay to verify the direct interaction between UBE2MP1 and miR-145-5p. As discovered, the luciferase signal in either HepG2 or Hep3B cells transfected with miR-145-5p mimics was decreased significantly, after the transfection of the UBE2MP1/pMIR/WT vector, in comparison with the control ones. On the contrary, the transfection of the UBE2MP1/pMIR/MUT vector did not induce signal changes with significance (Figure 4F). Thus, the findings convincingly illustrated the ceRNA effect of UBE2MP1 on sponging miR-145-5p.
Interestingly, we occasionally noticed that RGS3 is highly expressed in both HCC cell lines and HCC tissues (Figure 5A–5D). And notably, RGS3 shared a positively correlated with UBE2MP1 expression in the HCC tissues from our center, and the expression of RGS3 presented a remarkable decrease in HCC cells when UBE2MP1 was depleted (Figure 5E). We wondered if there exists a co-expression or regulation between UBE2MP1 and RGS3. On basis of this point, we tried to explore the role that miR-145-5p plays in this axis. The microcosm an online prediction software (https://www.microcosm.com/) indicated that miR-145-5p was an upstream regulator potentially interacting with the 3’-untranslated region (3’-UTR) of RGS3 mRNA with a high recommendation (Mfe: -28.2 kcal/mol) (Figure 5F). The dual-luciferase reporter assay was used for validating this assumpted direct binding between miR-145-5p and RGS3 mRNA. Similar to the above technique, we constructed the vectors containing a 202 bp sequence intercepted from the 3'UTR from the RGS3 mRNA (WT-UTR), and also the control luciferase vectors containing a mutated miR-145-5p targeting site (MUT-UTR). Both HepG2 and Hep3B cells were transfected with either the above two kinds of vectors and also mimicked by miR-145-5p. Here, taking HepG2 cells, for example, the miR-145-5p mimics (HepG2/miR-145-5p) significantly defected the luciferase signal of RGS3/pMIR/WT in comparison with the negative control (HepG2/NigmiR). The signal suppressive effect induced by miR-145-5p was abrogated in HepG2 cells transfected with a mutated miR-145-5p binding site (Figure 5G). And similar results were obtained from the Hep3B cell line too.
Since the verification of the post-transcriptional degeneration effect of miR-145-5p on RGS3 mRNA, we prompt that the phenotypes induced by UBE2MP1 in HCC are due to the modulation of the potential miR-145-5p/RGS3 axis. Take HepG2 cells as an example, firstly we lower the miR-145-5p expression in HepG2 cells treated with UBE2MP1 depletion again (Figure 6A). The suppressed cell proliferation was significantly recovered along with the escape of G0/G1 stage arrest (Figure 6B, 6C). Simultaneously, the number of apoptotic cells was decreased (Figure 6D). Secondly, we re-introduced RGS3, which had been down-regulated through UBE2MP1 depletion, in HepG2 cells. As expected, the knock-down of miR-145-5p led to an obvious increase in RGS3 expression at both mRNA and protein stages (Figure 7A). Interestingly, even though the expression of either UBE2MP1 or miR-145-5p was not impacted by re-introducing RGS3 in HCC cells, the cell proliferation and apoptosis statuses induced by UBE2MP1 depletion were effectively recovered RGS3 (Figure 7B–7F). Last but not least, we simultaneously obtained similar results from the rescue experiments in Hep3B cell lines (Supplementary Figures 2, 3). All these findings above present an HCC-promoting axis of miR-145-5p/RGS3 under the control of UBE2MP1 through its ceRNA effect.
HCC composes the majority portion of liver cancers and results in unsatisfactory outcomes and high mortality, due to the reckless growth and invasiveness, and potent tumoral heterogeneity [15, 16]. Recent strategies for targeting the particular genes in treating HCC patients systemically provide the researchers with a mighty approach to conquer the bottleneck of the rapid progress, high recurrence, and shorten overall survival of HCC. However, the therapeutic progress made so far is far from enough, and the comprehensive intracellular mechanisms promoting HCC leave us a lot for illustrating. It has been a long time since the pseudogenes were regarded as ‘molecular fossils’ or ‘relics of evolution’ without definite intracellular functions [17]. However, anomalous transcription of the pseudogenes could be activated in the pathological situation and exerts particular effects on either oncogenic or tumor-suppressive events intracellular through different mechanisms [18, 19]. Recent evidence has gradually revealed the potential functions of pseudogene-derived transcripts, playing complex roles in the processes of transcription and post-transcription regulations and tumorigenesis and development [20]. In this study, we found that UBE2MP1 was activated to transcribe in HCC tissues, and mostly maintained at a high level, along with the common expression in the HCC cell lines much higher than the LO2 cells. UBE2M is the parental gene of UBE2MP1, and this gene encodes the protein as one of the key operators of the E2 ubiquitin-conjugating enzyme family [21]. As acknowledged, the E2 family is essential for the ubiquitination cascade, and UBE2M is a promoter of tumor onset and development, up-regulated in multiple human malignancies, including osteosarcoma, cholangiocarcinoma, and HCC [22, 23]. According to the detection of both the TCGA database and the real patients’ specimens, we validated the over-expression of UBE2M in HCC. Since the parental gene UBE2MP1 is a key mediator of ubiquitination, and has been reported as a tumor-promoter in HCC, we focused on this pseudogene and wondered if the abnormally transcribed UBE2MP1 participates in HCC cell growth dependently or independently to UBE2MP1. However, as a pseudogene of UBE2M, barely transcribed in the liver, there is no adequate information and report of its functions in cancer research. The analysis of the real patients’ clinicopathological features demonstrated that the expression of UBE2MP1 is correlated with severe tumor progress and dismal phenotypes hindering a better prognosis and overall survival. We suppose that UBE2MP1 transcription participates in HCC progress through some mechanisms dependent or independent of its parental gene UBE2M. Based on the expression modulation of UBE2MP1 in HCC cell lines without inducing significant changes on UBE2M, we observed an efficient influence of either cell growth and maintenance when the UBE2MP1 transcript was depleted. The abrogation of UBE2MP1 not only suppressed tumor cell growth by arresting the cell cycle in the early stages, but also induced cell apoptosis. Obviously, the in vitro experiment indicated the independent effects of UBE2MP1 in the facilitation of HCC progress. Wondering the exact mechanism of UBE2MP1 in promoting HCC growth, we investigated the possibility of the ceRNA effect implemented by its pseudogene-derived transcript. MiR-145-5p has been reported mainly to play the tumor-suppressive role in multiple malignancies. For instance, miR-145-5p induces tumor cell apoptosis in prostate cancer by degrading WIP1 [24]; MiR-145-5p is extremely decreased in breast cancer and attenuates paclitaxel resistance and suppresses the progression of breast cancer cells by targeting SOX2 [25]. For liver cancer, recent reports have noticed the tumor-suppressive effects of miR-145-5p in either cholangiocarcinoma or HCC by targeting different mRNAs, like CDCA3, and SPATS2, and involved in cell proliferation, apoptosis, or metastasis [26, 27]. Thus, we suppose that miR-145-5p is a key pivot of multiple regulating axes, impacting HCC progress and development through different mechanisms. RGS3 gene encodes a regulator of the G-protein signaling (RGS) family and functions as a GTPase-activating protein for inhibiting G-protein-mediated signal transduction [28]. Accumulating evidence has indicated the pivotal function of RGS3 in participating in the Wnt signaling and the epithelial-mesenchymal transition, which strongly suggested RGS3 as a tumor enhancer. For example, RGS3 is highly expressed in gastric cancer and efficiently promotes tumor growth, and is correlated with poor prognosis [29]. And recently, RGS3 was reported to hinder the effect of KRASG12C inhibitors in the targeted therapy of lung cancer by enhancing the GTPase activity of KRAS [30]. In gastric cancer, RGS3 has been reported overexpressed in tumor cells and played a critical role in the Wnt signaling pathway on epithelial-mesenchymal transition [31]. For HCC, the limited reports suggested an enhancement of HCC cell apoptosis after its indirect down-regulation [32]. However, the definite function of RGS3 and the relative mechanism in HCC is not sufficient yet. As mentioned in the introduction section, pseudogene transcripts may exert the ceRNA effect in tumor progression. For instance, the transcript of pseudogene BRAFP1 promotes lymphoma development by sponging a series of miRNAs, like miR-134 and miR-653, and efficiently preserves the expression of its parental gene BRAF and leads to the activation of the downstream MAPK pathway [33]; On the contrary, pseudogene PTENP1 exerts the ceRNA effects on the oncogenic miR-21 and miR-19, and sequentially protects the expression of its parental gene PTEN and assists to suppressive tumor development in gastric cancer and clear cell renal carcinoma [34, 35]. Simultaneously, Our recent study illustrated a ceRNA effect of pseudogene AKR1B10P1 transcript, sponging the tumor-suppressive miR-138 in HCC and resulting in an enhancement of tumor growth [36]. Similarly, in this study, we predicted the direct interaction of miR-145-5p with both RGS3 and the UBE2MP1 transcript at the same time. MiR-145-5p is a pivotal post-transcriptional regulator that has been described as a tumor suppressor in multiple human cancers. As reported, miR-145-5p was decreased to a low level in colon cancer, and elevation of miR-145-5p expression could significantly arrest the tumor cell cycle at the G0/G1 phase [37]. In liver cancer, miR-145-5p was reported as an inhibitor of cell proliferation by negatively targeting the oncogene Spermatogenesis associated serine-rich 2 (SPATS2) and Kruppel-like factor 5 (KLF5) [27, 38]. However, the further mechanism of miR-145-5p on HCC is in limitation, and we wonder if miR-145-5p is present to be the critical bridge between RGS3 and UBE2MP1. From this point, we separately discussed the existence of the degrading effect of miR-145-5p on RGS3 mRNA and the ceRNA effect between UBE2MP1 and miR-145-5p. Firstly, the typical post-transcriptional regulation of miR-145-5p was validated through the dual-luciferase reporter assay and the 3’-UTR of RGS3 mRNA provides a particular site for binding to miR-145-5p. Since the complex network between the miRNAs and the verity of downstream mRNAs, we thought that the miR-145-5p/RGS3 axis might be one of the functional pathways affecting HCC progress and development. On the other hand, we discovered a specific binding site on the sequence of UBE2MP1 transcript for sponging miR-145-5p, and a remarkable increase of miR-145-5p in HCC cell lines through depleting UBE2MP1 supported the ceRNA effect of UBE2MP1 on defecting miR-145-5p and the tumor-suppressive effect. Moreover, the positive correlation between UBE2MP1 transcript and RGS3, and also the consequential strong decline of RGS3 through either UBE2MP1 depletion or elevation of miR-145-5p, definitely demonstrate RGS3 as a downstream effector controlled by UBE2MP1. Simultaneously, the impairment of cell growth and apoptosis resistance induced by UBE2MP1 depletion was significantly recovered by knocking down miR-145-5p or re-introducing RGS3. Thus, all these findings provide a convincible validation of the enhancement of HCC growth and maintenance of resisting cell apoptosis under the control of pseudogene UBE2MP1 through modulating the miR-145-5p/RGS3 axis. In summary, our study discovered the transcriptional activation of pseudogene UBE2MP1 in HCC and validated the adverse correlation between UBE2MP1 transcript and outcomes in HCC patients. As the in vitro experiment demonstrated, the UBE2MP1 transcript potently facilitates HCC cell proliferation and apoptosis resistance, and the ceRNA effect of it provides a critical mechanism of UBE2MP1, independent of its parental gene, in facilitating HCC cell growth and maintenance via modulating the miR-145-5p/RGS3 axis. However, limited by lacking the in vivo experiment, it still needs further exploration of the exact function of the UBE2MP1 transcript in the animal models, and we intend to carry out the orthotopic transplantation of mouse liver to give out more convincible evidence to support our findings. And lastly, we suppose that UBE2MP1 and its downstream miR-145-5p/RGS3 axis might be the potential and hopeful targets for HCC prevention and therapeutic strategy, even though there is still much detail for investigation. | true | true | true |
PMC9596212 | 36205565 | Zhipeng Sun,Guangyang Chen,Liang Wang,Qing Sang,Guangzhong Xu,Nengwei Zhang | APEX1 promotes the oncogenicity of hepatocellular carcinoma via regulation of MAP2K6 | 04-10-2022 | APEX1,HCC,MAP2K6,tumor growth | Objective: Apurinic/apyrimidinic endonuclease 1 (APEX1), a key enzyme responsible for DNA base excision repair, has been linked to development and progression of cancers. In this work, we aimed to explore the role of APEX1 in hepatocellular carcinoma (HCC) and elucidate its molecular mechanism. Methods: The expression of APEX1 in HCC tissues and matched adjacent normal tissues (n = 80 cases) was evaluated by immunohistochemistry. Web-based tools UALCAN and the Kaplan-Meier plotter were used to analyze the Cancer Genome Atlas database to compare expression of APEX1 mRNA to 5-year overall survival. APEX1 was stably silenced in two HCC cell lines, Hep 3B and Bel-7402, with shRNA technology. An in vivo tumorigenesis model was established by subcutaneously injecting sh-APEX1-transfected Bel-7402 cells into mice, and tumor growth was determined. We performed high-throughput transcriptome sequencing in sh-APEX1-treated HCC cells to identify the key KEGG signaling pathways induced by silencing of APEX1. Results: APEX1 was significantly upregulated and predicted poor clinical overall survival in HCC patients. Silencing APEX1 inhibited the proliferation of HCC cells in vivo and in vitro, and it repressed invasion and migration and increased apoptosis and the percentage of cells in G1. Differentially expressed genes upon APEX1 silencing included genes involved in TNF signaling. A positive correlation between the expression of APEX1 and MAP2K6 was noted, and overexpressing MAP2K6 overcame cancer-related phenotypes associated with APEX1 silencing. Conclusion: APEX1 enhances the malignant properties of HCC via MAP2K6. APEX1 may represent a valuable prognostic biomarker and therapeutic target in HCC. | APEX1 promotes the oncogenicity of hepatocellular carcinoma via regulation of MAP2K6
Objective: Apurinic/apyrimidinic endonuclease 1 (APEX1), a key enzyme responsible for DNA base excision repair, has been linked to development and progression of cancers. In this work, we aimed to explore the role of APEX1 in hepatocellular carcinoma (HCC) and elucidate its molecular mechanism. Methods: The expression of APEX1 in HCC tissues and matched adjacent normal tissues (n = 80 cases) was evaluated by immunohistochemistry. Web-based tools UALCAN and the Kaplan-Meier plotter were used to analyze the Cancer Genome Atlas database to compare expression of APEX1 mRNA to 5-year overall survival. APEX1 was stably silenced in two HCC cell lines, Hep 3B and Bel-7402, with shRNA technology. An in vivo tumorigenesis model was established by subcutaneously injecting sh-APEX1-transfected Bel-7402 cells into mice, and tumor growth was determined. We performed high-throughput transcriptome sequencing in sh-APEX1-treated HCC cells to identify the key KEGG signaling pathways induced by silencing of APEX1. Results: APEX1 was significantly upregulated and predicted poor clinical overall survival in HCC patients. Silencing APEX1 inhibited the proliferation of HCC cells in vivo and in vitro, and it repressed invasion and migration and increased apoptosis and the percentage of cells in G1. Differentially expressed genes upon APEX1 silencing included genes involved in TNF signaling. A positive correlation between the expression of APEX1 and MAP2K6 was noted, and overexpressing MAP2K6 overcame cancer-related phenotypes associated with APEX1 silencing. Conclusion: APEX1 enhances the malignant properties of HCC via MAP2K6. APEX1 may represent a valuable prognostic biomarker and therapeutic target in HCC.
Hepatocellular carcinoma (HCC) was the most common type of primary liver cancer in adults and the sixth most commonly diagnosed cancer in 2018 [1]. Due to a limited array of treatment options, poor prognoses, and high recurrence rates, HCC has become one of the most fatal cancers in the world [2]. In recent years, however, the diagnosis and treatment of HCC have been greatly improved with methods such as surgical resection, liver transplantation, and various local treatments. Although these treatments have helped improve efficacy, the prognosis for HCC patients remains poor [3, 4]. Progress in this realm is limited by a lack of clarity regarding the exact mechanism of pathogenesis of HCC, which is known to involve multiple genetic changes. Our present understanding of the mechanism of pathogenesis of HCC includes gene mutations, genetic changes to metabolism, abnormalities of intracellular signaling pathways, and changes to the local tumor microenvironment, with gene mutations being the most complex [5]. After certain key gene mutations occur, interactions among a series of signaling proteins are altered, potentially leading to unregulated proliferation, metastasis, and other changes of hepatocytes [6]. Because of the importance of gene mutations in mechanisms underlying cancer biology, targeted gene therapy has emerged as an attractive area of study in the field of cancer treatment in recent years. In addition, the identification of more tumor biomarkers can potentially aid in the early diagnosis of disease and serve as a method to enable monitoring of the efficacy of treatment [7–9]. Therefore, it is particularly important to identify key pathogenic genes associated with HCC. Apurinic/apyrimidinic endonuclease 1 (APEX1) is crucial to DNA base excision repair [10, 11]. It acts via oxidation-reduction interactions and can regulate the DNA-binding activity of various transcription factors [12]. It is also known as an oxidation-reduction factor (Ref-1) and is an important multifunctional protein in the human body. APEX1 participates in various cell reactions such as cell proliferation, apoptosis, and differentiation by regulating the activity of oxidation-reduction-sensitive transcription factors [13, 14]. APEX1 is generally expressed in a variety of human cells at a high level. A study conducted by Cao L et al. showed that APEX1 is up-regulated in HCC and that this over-expression correlates with cancer aggressiveness [15]. In HCC, oxidative injury is important to the process of carcinogenesis, which promotes the development of tumors in many ways, including cell apoptosis and resistance against cell death signaling pathways [16]. Mitogen-activated protein kinase kinase 6 (MAP2K6) is an upstream kinase of the TNF signaling pathway, which is involved in various physiological and pathological processes including cell growth, development, division, and inflammatory reactions [17–19]. Guo Y et al. discovered that MAP2K6 enhanced the sensitiveness of paclitaxel for ovarian cancer via inducing autophagy [20]. At present, it is believed that MAP2K6 may be associated with the occurrence and progression of tumors and could be potentially treated as a new diagnostic or prognostic biomarker for cancers. In this study, we determined the expression of APEX1 in human HCC tissues. Lentivirus mediated RNA interference was used to silence APEX1. The influence of APEX1 gene expression on the biological behavior of HCC cells was explored in vivo and in vitro, and we explored possible mechanisms of pathogenesis.
Eighty pairs of HCC/paracancerous tissues were collected from HCC patients who underwent surgical treatment in the oncology surgery department, Beijing Shijitan Hospital, Capital Medical University (Peking University Ninth School of Clinical Medicine) from August 2015 to October 2018. All patients did not receive preoperative radiotherapy or chemotherapy, and the adjacent tissues were at least 5 cm away from the edge of the tumor. All fresh tissue samples were immediately stored in a liquid nitrogen tank to protect RNA from degradation. The use of human tissues was approved by the Ethics Committee of Beijing Shijitan Hospital, Capital Medical University (NO: sjtky11-1x-2020(15)) and was conducted in accordance with the Declaration of Helsinki. We obtained written informed consent from every patient. The animal experimental protocol was approved by the Animal Care Committee of Beijing Shijitan Hospital, Capital Medical University (NO: sjtky11-1x-2019(28), sjtky11-1x-2018(108) and sjtky-1x-2019(89)).
HCC cell lines (Huh-7, SMMC-7721, Hep G2, Hep 3B, HCC-9204, Bel-7402, and Bel-7405) and the normal liver cell line L-02 were obtained from the National Biomedical Experimental Cell Resource Bank (Beijing, China). Cells were maintained in RPMI 1640 supplemented with 10% FBS (Gibco, Grand Island, NY, USA), 100 U/mL penicillin and 100 μg/mL streptomycin (Gibco, Grand Island, NY, USA). Then, the mixture was cultured in a 37°C incubator with a 5% CO2 environment. When the cells reached approximately 85% confluence, they were passaged at a ratio of 1:3.
Specific shRNA strands (F: 5′ CCG GCA GAG AAA TCT GCA TTC TAT TCT CGA GAA TAG AAT GCA GAT TTC TCT GTT TTT, R: 5′ AAT TCA AAA ACA GAG AAA TCT GCA TTC TAT TCT CGA GAA TAG AAT GCA GAT TTC TCT G) were obtained from Sangon Biotech (Shanghai, China). These shRNA were used to construct plasmid pLKO.1-puro-shRNA. A corresponding control plasmid (pLKO.1-puro-Ctrl) was also constructed. APEX1-knockdown lentiviruses were created with plasmids psPAX2 and pMD2.G (BIOFENG, Shanghai, China) and were transfected into 293T cells with pLKO.1 puro-shRNA using Lipofectamine 2000 Transfection Reagent (Thermo Fisher Scientific, MA, USA). Hep 3B and Bel-7402 cells were then infected with these lentiviruses at a multiplicity of infection of 10 and 20, respectively, and selected with puromycin (2 μg/mL) (Thermo Fisher Scientific, MA, USA). At 2 weeks post-infection, cells were harvested, and the protein expression levels of APEX1 were determined.
All tissue samples were fixed in 10% neutral formalin, embedded in paraffin, sectioned, dewaxed with xylene and dehydrated with gradient ethanol. Then, citric acid buffer was used for high temperature antigen repair, 3% hydrogen peroxide (Solarbio, Beijing, China) was used to block endogenous peroxidase activity, and the samples were incubated at room temperature for 20 min. Next, the slides were incubated with an anti-APEX1 primary antibody (Abcam, Cambridge, UK) at 4°C overnight. The next day, the samples were incubated with an appropriate horseradish peroxidase-conjugated secondary antibody at 37°C for 30 min. After staining with 3, 3′-diaminobenzidine (Solarbio, Beijing, China), the samples were stained with hematoxylin, dehydrated with gradient ethanol, made transparent with xylene, and sealed with neutral gum. The integrated option density (IOD) of APEX1 was chosen to determine the semiquantitative protein expression. ImageJ software (version 1.2; WS Rasband, National Institute of Health, Bethesda, MD, USA) was used to conduct deconvolution and downstream analyses.
Cells were collected, and total protein was extracted from cell lysates. SDS-PAGE electrophoresis was performed, and proteins were transferred electrophoretically from the gel to a polyvinylidene fluoride (PVDF) membrane. Tris-buffered saline liquid containing Tween 20 (TBST) and 5% skimmed milk powder was used to block the PVDF membrane for 2 h at room temperature. The primary anti-APEX1 antibody (Abcam, Cambridge, UK), diluted in 1% skimmed milk powder in TBST, was added, and the membrane was incubated on a shaker at 4°C overnight. The secondary antibody was added for 2 h at room temperature. Enhanced chemiluminescence reagent was added to the membrane, which was then developed in a chemiluminescence imaging system instrument.
Total RNA was extracted from cells by the Trizol method, and cDNA was synthesized according to the instructions of the manufacturer of the M-mlv reverse transcriptase kit (Takara Bio, Beijing, China). PCR amplification was performed with EvaGreen Dye (Biotium, Fremont, CA, USA) as follows: 95°C for 5 min followed by 40 cycles of 95°C for 20 s, 60°C for 30 s, and 72°C for 20 s. The primers for APEX1 were F: 5′-GCT GCC TGG ACT CTC TCA TCA AT-3′ and R: 5′-CCT CAT CGC CTA TGC CGT AAG AA-3′. The primers for MAP2K6 were F: 5′-TGT GCA TTT CCA TCT TGA TTC CC-3′ and R: 5′-CGC TTC TTG CCT TTC GAC TG-3′. The primers for TNFAIP3 were F: 5′-CTG CCA GCG AGC GAG C-3′ and R: 5′-GTG CTC TCC AAC ACC TCT CC-3′. The primers for CASP3 were F: 5′-TGG AAC CAA AGA TCA TAC ATG GAA-3′ and R: 5′-TTC CCT GAG GTT TGC TGC AT-3′. The primers for the internal control GAPDH were F: 5′-TGA AGG TCG GAG TCA ACG G-3′ and R: 5′-TCC TGG AAG ATG GTG ATG GG-3′.
Cell viability was measured by the Cell Counting Kit 8 (CCK8) in accordance with the manufacturer’s instructions (Solarbio, Beijing, China). The prepared cells were collected and resuspended in 1640 medium (10% FBS) at 2 × 104/mL. Then, 0.1 mL of this suspension was placed into wells of a 96-well plate and cultured for 0 h, 24 h, 48 h, or 72 h. Cell proliferation rates were quantified by measuring OD (450 nm) values with a microplate reader. For colony formation assays, treated were collected and added to a 6-well plate at a density of 1000 cells per well. The cells were cultured normally until colonies were visible. The cell colonies were stained with 0.1% crystal violet solution and photographed with a camera, then the number of cell colonies in each group was counted manually. The EdU assay was performed with an EdU kit (Roche, Indianapolis, IN, USA) according to the manufacturer’s instructions. Results were analyzed with a flow cytometer equipped with CellQuest software (BD Biosciences, San Diego, CA, USA).
HCC cells were plated and grown to 70% confluence on 6-well plates and were wounded with 1-mL pipette tips. Samples were examined at 0 and 48 h after scratching, and the wound healing status of each group was observed and photographed.
Detection of cell invasion was performed in 8-μm Transwell chambers (BD Biosciences, USA). Transfected Hep 3B and Huh7 cells were suspended in serum-free medium and then seeded into the Matrigel-upper chambers at 5 × 104 cells per well. DMEM containing 10% FBS (500 μL) was added into the lower chamber. After culturing for 24 h, the invasive cells were stained with 0.1% crystal violet and photographed under a light microscope. Cell numbers was counted manually.
Transfected Hep 3B and Huh7 cells were processed with an Apoptosis Detection Kit (Beyotime, China), according to the manufacturer’s instructions. Briefly, transfected cells were washed twice with cold PBS. After incubation with 5 μL of FITC-Annexin V and 5 μL propidium iodide for 20 min in the dark, apoptosis of Hep 3B and Huh7 was detected using a flow cytometer (FACScan; BD Biosciences, USA). For cell cycle analyses, cells were collected and incubated with 70% ethanol at 4°C overnight for fixation. The cells were washed twice with PBS and incubated with 100 μg/mL RNase A and 50 μg/mL propidium iodide for 1 h at 37°C. The percentage of cells in each phase of the cell cycle was then measured by flow cytometry (FACScan; BD Biosciences, USA).
Total RNA was extracted from cells that had been transfected with shRNA targeting APEX1 (sh-APEX1) or a negative control (sh-NC), and reverse transcription was performed to construct a cDNA library. High-throughput transcriptome sequencing was performed with an Illumina Hiseq 2500 system, and transcriptome sequencing data were obtained for bioinformatics analysis.
BALB/c nude mice (male, 6–8 weeks) were adaptively raised for 2 weeks under SPF conditions, and the mice were randomly divided into two groups at the third week. There were 10 mice in each group, and each mouse was numbered. The right armpit of nude mice was subcutaneously injected with 200 μL of a 1 × 107/ml cell suspension. After 8 weeks post-inoculation, the mice were killed by dislocation, and the tumor rate was calculated. The tumor tissue was completely separated, the mass of the tumor was measured, and the tumor volume was calculated using the equation: tumor volume (mm3) = 1/2 (long diameter × short diameter × short diameter). For the Establishment of the orthotopic liver cancer model, the mice were anesthetized with 3% to 5% isoflurane, and the left lobe of the liver was exposed along the midline of the upper abdomen. Suspended cells are sucked into the syringe, punctured into the left lobe of the liver about 5 mm at an angle of 30°, and the cells are slowly injected into the liver tissue and then multi-point injections are performed. Then, gently press the injection site with a cotton ball for about 1 minute to reduce bleeding and cell suspension leakage. Finally, the peritoneum was closed with 4–0 sutures intermittently, and the skin incision was closed with suture nails. After execution, liver tissue was taken for analysis of tumor formation. The animal experimental protocol was approved by the Animal Care Committee of Beijing Shijitan Hospital, Capital Medical University (NO: sjtky11-1x-2019(28), sjtky11-1x-2018(108) and sjtky-1x-2019(89)).
All statistical analyses were performed in Statistical Product and Service Solutions (SPSS) 20.0 software (SPSS, Chicago, IL, USA). SigmaPlot12.3 (Systat Software, San Jose, CA, USA) and GraphPad Prism 5.0 (GraphPad Software, La Jolla, CA, USA) software were used to draw graphs. Student’s t-test, one-way analysis of variance (ANOVA), and a rank-sum test were flexibly applied according to the conditions at hand. Differences for which P < 0.05 were regarded as statistically significant.
Immunohistochemistry was used to measure the expression of APEX1 in HCC tissues and adjacent normal tissues (n = 80). As shown in Figure 1A, expression of APEX1 protein in HCC tissues was higher relative to that in adjacent normal tissues. Furthermore, we measured the expression of APEX1 mRNA using UALCAN, an online server derived from the Cancer Genome Atlas (TCGA) dataset [17]. The results indicated that expression of APEX1 mRNA in HCC tissues at multiple stages was higher than that in normal tissues (Figure 1B). Western blotting was used to measure the expression of APEX1 protein in HCC cell lines (Huh-7, SMMC-7721, Hep G2, Hep 3B, HCC-9204, Bel-7402, and Bel-7405) and in the normal liver cell line L-02. As shown in Figure 1C, APEX1 expression in the HCC cell lines was higher, to varying degrees, than was expression in L-02.
The relationships between APEX1 expression and clinical characteristics of HCC patients are shown in Table 1. Statistical analyses suggested that APEX1 expression level increased with increasing pathological grade and TNM stage of HCC (P < 0.05). Moreover, the Kaplan-Meier plotter [18], a web-based tool that uses TCGA databases to assess the effect of changes to the expression of 54,000 genes (at the level of mRNA, miRNA, or protein) on survival, was used. This plotter allows the prediction of the prognostic value of the tumor-specific expression of various genes. When we analyzed the correlation of APEX1 expression with HCC patient survival in this way, the results suggested that patients with high APEX1 expression exhibited a lower 5-year survival rate compared with those with low APEX1 expression (Figure 2).
To identify functional roles of APEX1 protein in HCC cells, we silenced the APEX1 gene in the two HCC cell lines, Hep 3B and Bel-7402, that were determined to have the highest expression of APEX1. The results of qRT-PCR and Western blot analyses suggested that the expression of APEX1 in the sh-APEX1 group was markedly lower than in the sh-NC group (Figure 3A, 3B). We next evaluated the vitality of these two cell lines by performing CCK-8 (Figure 3C), colony-forming (Figure 3D) and EdU assays (Figure 3E). These results indicated that APEX1 knockdown suppressed the proliferation of HCC cells. We also examined the effects of APEX1 on the invasion and migration of HCC cells. Inhibition of APEX1 expression appeared to decrease the invasiveness of the HCC cells (Figure 3F). Similarly, APEX1 knockdown attenuated the migration ability of HCC cells (Figure 3G). Moreover, we applied flow cytometry to detect whether the anti-proliferation activity of APEX1 is associated with cell apoptosis or the cell cycle. The results showed a significant increase in cell apoptosis upon APEX1 knockdown (Figure 3H). Analysis of the cell cycle distribution revealed that the percentage of cells in G1 was increased upon the inhibition of APEX1 in both cell lines (Figure 3I). Taken together, these data indicated that APEX1 promotes HCC cell proliferation, invasion, and migration, and it inhibits cell apoptosis and alters cell cycle distribution of HCC cells.
To investigate the role of APEX1 in HCC tumor growth in vivo, we established a tumorigenesis model by subcutaneously injecting the sh-APEX1-transfected HCC cell line Bel-7402 into a nude mouse model. Tumor volume was monitored every week, and mice were sacrificed 8 weeks after inoculation for the determination of tumor weight. Compared with the control group, implantation of cells under-expressing APEX1 caused a significantly increased tumor growth in terms of tumor volume and weight (Figure 4A, 4B). Results of immunohistochemical analyses of isolated tumors confirmed a decreased expression of APEX1 protein in the sh-APEX1 group (Figure 4C). In addition, the establishment of the orthotopic liver cancer model revealed that APEX1 knockdown suppressed numbers of tumor nodules (Figure 4D). These findings indicate that silencing of APEX1 can reduce the tumorigenicity of HCC cells in vivo, thereby potentially inhibiting the development of HCC.
In order to validate the transcriptome changes in sh-APEX1-transfected HCC cells, we performed high-throughput transcriptome sequencing. The stratified cluster heat map demonstrated the distribution of differential genes between the sh-APEX1 group and the control group. There were 84 up-regulated genes and 39 down-regulated genes in the sh-APEX1-treated HCC cells (Figure 5A). A total of 123 differentially expressed genes (DEGs) were subjected to Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathway analysis. The molecular functions that the DEGs were found to be involved in included binding, catalytic activity, transporter activity, molecular function regulation, structural molecule activity, nucleic acid binding transcription factor activity and molecular transducer activity. The cellular components of the gene products included cell, cell part, organelle, membrane, organelle part, membrane part, extracellular region, extracellular region part, and synapse. The biological processes involved included cellular process, single-organism process, biological regulation, regulation of biological process, metabolic process, multicellular organismal process, and response to stimulus and signaling (Figure 5B). KEGG enrichment analysis showed that after silencing of APEX1, the top 20 signaling pathways that DEGs are involved in were related to amyotrophic lateral sclerosis (ALS); the synaptic vesicle cycle; proximal tubule bicarbonate reclamation; glycine, serine and threonine metabolism; neomycin, kanamycin and gentamicin biosynthesis; carbohydrate digestion and absorption; transcriptional misregulation in cancers; the tumor necrosis factor (TNF) signaling pathway; endocrine and other factor-regulated calcium reabsorption and arginine and proline metabolism (Figure 5C). Results of a differential gene protein interaction network analysis showed that APEX1, CASP3, EGR1, IGFBP3, INHBA, TNC, HOXA10, IGFBP3, MMP23, HMGA2, DAPK2 and TXNIP genes made up the core protein interaction network (Figure 5D). We next performed a qRT-PCR analysis to evaluate the expression of 3 DGEs that are enriched in the TNF signaling pathway: those encoding mitogen-activated protein kinase kinase 6 (MAP2K6), TNF alpha-induced protein 3 (TNFAIP3) and caspase 3 (CASP3). According to this assay, the expression of MAP2K6 and TNFAIP3 were evidently decreased, while CASP3 expression was significantly increased, in sh-APEX1-treated HCC cells as compared to control-treated cells (Figure 5E).
Based on the down-regulation of MAP2K6 in sh-APEX1-treated HCC cells and a positive expression correlation between APEX1 and MAP2K6 in HCC tissues (Figure 6A), we predicted that APEX1 enhances the malignant properties of HCC via its impact on MAP2K6 expression. We therefore sought to investigate whether MAP2K6 was responsible for the declined proliferation, invasion, and migration, enhanced cell apoptosis and altered cell cycle distribution induced by silencing of APEX1. We enhanced the expression of MAP2K6 in sh-APEX1-treated HCC cells and found that overexpression of MAP2K6 could abolish the role of inhibition of APEX1 in attenuating cell vitality, colony formation and proliferation (Figure 6B–6D). The decreases in cell invasion and migration induced by sh-APEX1 were also reversed by MAP2K6 overexpression (Figure 6E, 6F). In addition, the significant increase in cell apoptosis and cells in the G1 phase caused by APEX1 knockdown were also attenuated by MAP2K6 expression (Figure 6G, 6H). These data indicate that MAP2K6 can inhibit the anti-tumorigenic function of sh-APEX1 in HCC cells.
Hepatocellular carcinoma is the most common type of primary liver cancer in adults [1, 2]. Although exact details of its mechanism of pathogenesis are still unclear, the malignant transformation of liver cells involves multiple genetic aberrations. At present, the main mechanisms of pathogenesis are thought to include gene mutations as well as changes to metabolism, intracellular signaling pathways and the local tumor microenvironment [19]. After some key gene mutations, a cascade of changes to signaling proteins occurs, and these changes then lead to the infinite proliferation and other behavioral changes. With the rapid development of genome sequencing technology in recent years, it has been found that some common gene mutations in liver cancer can also cause changes in other cells types. Increasing attention has thus been devoted toward targeted gene therapy as a method of cancer treatment [20]. At the same time, tumor markers can aid in effective early diagnosis of disease, and can also be a tool for monitoring responses to therapy. Therefore, it is particularly important to identify key pathogenic genes [21, 22]. APEX1 is a multifunctional protein that is widely expressed in various types of human cells. It is encoded by an important gene [10, 12]. APEX1 has two key roles. It is not only responsible for repairing the apurinic/apyrimidinic (AP) sites generated by various factors and thus maintaining the stability of genomic DNA, but it is also responsible for regulating DNA binding activity by controlling the activation state of transcription factors [12–14]. Whereas the genomic instability that results from improperly repair AP sites typically results in controlled cellular self-destruction, in the presence of abnormal binding of transcription factors to DNA, the cell may undergo malignant transformation and develop into a tumor. It has been shown that there are significantly higher levels of APEX1 protein in tumor cells compared to normal cells [15, 23, 24]. In addition, Di Maso et al. reported that the expression of APEX1 mRNA was significantly increased in HCC tissues [25]. In our study, we measured the expression of APEX1 protein and mRNA in HCC tissues and cells and in normal liver tissues and cells. Indeed, we found that expression of the APEX1 gene was significantly up-regulated in HCC. Moreover, APEX1 expression was found to be closely associated with the pathological grade of HCC: highly expressed APEX1 predicted poor clinical overall survival in HCC patients. This indicates that APEX1 is up-regulated in HCC and that it may contribute to the development of HCC. Aberrant expression of APEX1 has been frequently identified in cancers and plays crucial roles in the modulation of multiple oncogenic properties. For example, APEX1 expression is inversely associated with survival among patients with breast cancer [26], gastric cancer [27] and prostate cancer [28]. In these studies, APEX1 was shown to be essential in modulating the growth of the malignancies. These findings are consistent with those reported here and demonstrate that APEX1 plays a crucial role in the tumorigenesis of multiple cancers. The knockdown of APEX1 repressed proliferation, invasion, and migration, accelerated cell apoptosis, and the percentage of cells in the G1 phase of the cell cycle of HCC-derived cells. To further verify those effects, we performed a tumor study in nude mice. We found that silencing of APEX1 markedly reduced the body weight and tumor volume of nude mice, which indicated that sh-APEX1 could reduce the tumorigenic characteristics of HCC cells. Next, to explore the molecular mechanism that would explain the role of APEX1 in the development of HCC, we screened for differentially expressed genes in HCC cells with knocked down APEX1 expression. We then identified the KEGG signaling pathways induced by the silencing of APEX1. KEGG [29] enrichment analysis indicated that DEGs were mainly involved in amyotrophic lateral sclerosis, the synaptic vesicle cycle, proximal tubule bicarbonate reclamation, glycine biosynthesis, and the TNF signaling pathway. Three TNF pathway-related DEGs, MAP2K6, TNFAIP3 and CASP3, were selected for analysis, due to their potential role as downstream regulatory genes and signaling pathways related to APEX1. We found by qRT-PCR analysis that TNFAIP3 and MAP2K6 were down-regulated and CASP3 was up-regulated after silencing of APEX1. MAP2K6 plays an important role in the p38 MAP kinase signaling cascade, which regulates many stress-induced responses, is related to multiple pathological conditions and plays a role in many cellular processes, including differentiation, in bone, muscle, and adipose tissue [30, 31]. Based on the down-regulation of MAP2K6 in sh-APEX1-treated HCC cells and a positive correlation of expression of APEX1 and MAP2K6 in HCC tissues, we suggest that APEX1 enhances the malignant properties of HCC via MAP2K6. Thus, we enhanced the expression of MAP2K6 in APEX1-silenced HCC cells to evaluate the anti-oncogenic function of MAP2K6 with silenced APEX1. Here, we found that forced expression of MAP2K6 could abolish the role of inhibition of APEX1 in attenuating cell vitality, colony formation, proliferation, migration and invasion. In addition, a significant increase in the cell apoptosis and cells in the G1 phase that was caused by APEX1 knockdown were also attenuated by MAP2K6 overexpression. These data indicate that MAP2K6 can inhibit the anti-tumorigenic function of sh-APEX1 in HCC cells. The above evidence enhanced our understanding of the molecular mechanisms of the role of APEX1 in regulating the aggressiveness of HCC.
We revealed that APEX1 is an independent prognostic factor that promotes HCC growth and metastasis through its interaction with MAP2K6. We confirmed that the upregulation of APEX1 is a common phenomenon in HCC tissues and cell lines and is significantly correlated with the pathological grade and TNM stage of HCC. In addition, the novel HCC-related gene APEX1 enhances the malignant properties of HCC via the overexpression of MAP2K6. APEX1 may represent a valuable prognostic biomarker and therapeutic target for HCC. | true | true | true |
PMC9596241 | Anqi Duan,Hui Li,Wenlong Yu,Yongjie Zhang,Lei Yin | Long Noncoding RNA XIST Promotes Resistance to Lenvatinib in Hepatocellular Carcinoma Cells via Epigenetic Inhibition of NOD2 | 18-10-2022 | Background. Hepatocellular carcinoma (HCC) is a severe global health issue that still lacks of effective treatments. Lenvatinib is a novel tyrosine kinase inhibitor (TKI) that has been approved for the treatment of HCC. However, drug resistance is inevitable and limits the clinical application of lenvatinib. Till now, there is still little knowledge about the mechanisms under the resistance to lenvatinib in HCC. Long noncoding RNA (lncRNA) is a group of noncoding RNAs that play essential roles in various physiological activities including the chemoresistance. In the present study, through RNA sequencing, we discovered that lncRNA XIST was upregulated in HCC cells that was insensitive to lenvatinib. Mechanically, we found that lncXIST promotes lenvatinib resistance via activation of EZH2-NOD2-ERK axis in HCC cells. Our data suggest that targeting lncXIST/EZH2/NOD2/ERK axis might be a promising strategy to enhance the efficacy of lenvatinib against HCC cells. | Long Noncoding RNA XIST Promotes Resistance to Lenvatinib in Hepatocellular Carcinoma Cells via Epigenetic Inhibition of NOD2
Background. Hepatocellular carcinoma (HCC) is a severe global health issue that still lacks of effective treatments. Lenvatinib is a novel tyrosine kinase inhibitor (TKI) that has been approved for the treatment of HCC. However, drug resistance is inevitable and limits the clinical application of lenvatinib. Till now, there is still little knowledge about the mechanisms under the resistance to lenvatinib in HCC. Long noncoding RNA (lncRNA) is a group of noncoding RNAs that play essential roles in various physiological activities including the chemoresistance. In the present study, through RNA sequencing, we discovered that lncRNA XIST was upregulated in HCC cells that was insensitive to lenvatinib. Mechanically, we found that lncXIST promotes lenvatinib resistance via activation of EZH2-NOD2-ERK axis in HCC cells. Our data suggest that targeting lncXIST/EZH2/NOD2/ERK axis might be a promising strategy to enhance the efficacy of lenvatinib against HCC cells.
Hepatocellular carcinoma (HCC) is an aggressive cancer and ranks the third leading cause of cancer-related mortality [1]. It was estimated that there are over 900,000 new HCC cases and 800,000 deaths worldwide each year [2]. Many factors such as chronic viral hepatitis type B/C, excessive alcohol consumption, and exposure to aflatoxin can contribute to the HCC. Mechanically, various growth factors like vascular growth factor (PDGF), vascular endothelial growth factor (VEGF), and fibroblast growth factor (FGF) are involved in the progression of HCC [3]. Based on that, the Food and Drug Administration (FDA) approved the sorafenib, an oral tyrosine kinase inhibitor (TKI), for the treatment of HCC [4]. However, sorafenib showed limited clinical benefits in advanced HCC clinical treatment, and the overall 5-year survival rate is relatively low [4]. Lenvatinib was the second first-line drug that has been approved by the FDA for the treatment of HCC, due its noninferior survival benefit compared to sorafenib [5]. Lenvatinib acts mainly via inhibition of the angiogenesis in various solid cancers such as HCC, lung cancer, and thyroid cancer [6]. Although lenvatinib showed promising clinical values, the mechanisms underlying lenvatinib resistance are complicated and largely unknown. Thus, further investigations on the molecular basis of LR may provide novel insights into the identification of novel molecular targets to overcome it. Long noncoding RNA (lncRNA) are a heterogeneous group of noncoding RNAs (ncRNAs) with a transcript size of larger than 200 nt [7]. lncRNAs have been documented to play essential roles in various biological activities such as development, differentiation, and cell death. Till now, many lncRNAs have been found to regulate HCC cell response to sorafenib. For example, LINC01089 can contribute to sorafenib chemoresistance in HCC cells [8]. LINC01273 was also able to confer resistance to sorafenib in HCC cells [9]. However, there is still little knowledge about the role of lncRNA in regulating the HCC cell response to lenvatinib. In the current study, we identified lncXIST, which was upregulated in lenvatinib-resistant HCC cells. Further investigation found that lncXIST contributes resistance to lenvatinib via epigenetic inhibition of NOD2. Our findings suggest that lncXIST may be applied as a novel predictive biomarker and therapeutic target for lenvatinib resistance in HCC cells.
Human hepatocyte LO2 cells were obtained from Shanghai Aulu Biological Technology (Shanghai, China). HCC cells (Hep3B, HepG2, Huh7, and HA22T) were obtained from Shanghai Bank of Cell Culture (Shanghai, China). The lenvatinib-resistant HepG2 (HepG2/R) was created by stepwise escalation method: parental HepG2 cell was cultured with gradually increase doses of lenvatinib from 2 nM to 1 μM over 8 months. LO2 cells were cultured in DMEM, and HCC cells were cultured in RPMI1640 medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin and streptomycin in humidified air with 5% CO2 at 37°C. All cells were authenticated by STR profiling and tested for mycoplasma contamination by Shanghai Biowing Applied Biotechnology (Shanghai, China). SCH772984 and SB202190 were obtained from Selleck Chemicals (USA). The lenvatinib was obtained from MedChemExpress (USA), and all other routine chemicals were obtained from Sigma-Aldrich (USA).
Cell viability was measured by cell counting kit-8 (CCK-8) kit from Dojindo Lab (Japan). Briefly, 2000 cells were seeded into 96-well plates in triplicate, cultured overnight. After different treatments, cell viability was measured according to the manufacturer's guide. The results were measured at 450 nm by the microplate reader (BioTek, USA).
Cells were seeded into 6-well plate at the density of 1 × 105 cells/well and subjected to various treatments after adhering overnight. Apoptosis was measured by flow cytometry using the Annexin V-FITC Apoptosis Detection kit (BD Pharmingen, USA) according to the manufacturer's guide. Data was analyzed using FlowJo software.
Cells were seeded into 6-well plates at the density of 1 × 105 cells/well and transfected using the 5 μl Lipofectamine 2000 (Life Technologies, USA), which was mixed with 2 μl of 20 μM siRNA in 250 μl of Opi-MEM (Gibco, USA). The siRNAs/antisense oligonucleotides (ASOs) were obtained from GenePharma Ltd. (Suzhou, China).
Full length of lncXIST and of NOD2 cDNA was synthesized according to their coding sequence and cloned into the pcDNA3.1 vector. Plasmids were transfected into cells using the Lipofectamine 2000 according to the manufacturer's guide.
Total RNA was purified from the cells using the TRIzol (Life Technologies, USA) according to the manufacturer's guide. Real-time qPCR was conducted using the SYBR Green qPCR kit (Takara, Japan). The relative gene expression was calculated using the 2-ΔΔCt method and the samples were run in triplicate.
Total RNA was isolated from cells as described above. RNA sequencing was conducted using an Illumina HiSeq X-ten platform at Hangzhou HiBio Technology (China). In short, sequencing libraries were constructed using TruSeq Library Prep Kit (Illumina) and were sequenced on HisSeq7500 machine (Illumina). Reads were aligned to human hg19, and gene expression values were calculated via counting the reads mapping by the edgeR software. p values were adjusted using the Benjamini-Hochberg method for controlling the false discovery rate. Genes with an adjusted p value < 0.01 and fold change > 2 were considered differentially expressed.
ChIP assay was conducted using the ChIP kit (Abcam, USA) according to the manufacturer's guide. Briefly, cells were treated with formaldehyde and incubated for 10 min to generate DNA-protein complexes. Cell lysates were then sonicated to generate chromatin fragments of 200-300 bp and immunoprecipitated with EZH2 or H3K27me3-specific antibodies or IgG (negative control). Precipitated chromatin DNA was analyzed by qRT-PCR.
RIP was conducted to examine whether XIST could interact to bind with potential binding proteins. EZMagna RIP kit (Sigma-Aldrich) was used according to the manufacturer's guide. Briefly, cells were lysed and cellular extract was incubated with magnetic beads conjugated with antibodies that are specific against EZH2, SUZ12, or IgG for 6 h. Then, the beads were incubated with 0.1% SDS/0.5 mg/ml proteinase K for 0.5 h. The immunoprecipitated RNA was subjected to analysis by RT-PCR.
Total proteins were extracted from cells using the CHAPS lysis buffer (Beyotime, China). The quantification of protein was measured by the Bradford assay kit (Sigma). Equal amount of protein (10 μg) was loaded onto 10% SDS-PAGE and transferred to PVDF membrane (Millipore). The membrane was blocked with 5% skimmed milk for 1 h at room temperature. Then, the membrane was incubated with primary antibody at 4°C overnight. The following antibodies were used: phospho-ERK (cat: 4370; dilution: 1 : 1000; Cell Signaling Technology), ERK (cat: 4696; dilution: 1 : 1000; Cell Signaling Technology), phospho-p38 (cat: 4511; dilution: 1 : 1000; Cell Signaling Technology), p38 (cat: 8690; dilution: 1 : 1000; Cell Signaling Technology), phospho-FRS2 (cat: 3864; dilution: 1 : 1000; Cell Signaling Technology), FRS2 (cat: ab183492; dilution: 1 : 1000; Abcam), NOD2 (cat: ab31488; dilution: 1 : 1000; Abcam), EZH2 (cat: ab191250; dilution: 1 : 1000; Abcam), H3K27me3 (cat: ab6002; dilution: 1 : 1000; Abcam), and GAPDH (cat: ab8245; dilution: 1 : 5000; Abcam). Secondary antibodies conjugated to horseradish peroxidase were ordered from Sigma-Aldrich (USA). The results were visualized using Tanon™ High-sig ECL Western Blotting Substrate (Tanon, China) in Tanon-4600 instrument (Tanon, China). All experiments were repeated at least three times.
All experiments were independently repeated at least three times. GraphPad Prism 7.0 (GraphPad Software, La Jolla, CA, USA) was used for data processing and statistical analysis. Results are reported as the mean ± standard deviation (SD). Differences between the means of the groups were determined using one-way ANOVA with post hoc test. P < 0.05 was considered significant.
Firstly, we established lenvatinib-resistant HCC cells by exposing HepG2 cells to increasing doses of lenvatinib for over 9 months. As shown in Figure 1(a), lenvatinib-resistant HepG2 cells (HepG2/R) showed much higher viabilities than parental HepG2 cells after treated with various doses of lenvatinib for 24 h. HepG2 and HepG2/R cells were treated with lenvatinib (20 μM) for 24 h, the FGFR signalling pathway was examined. It was found that lenvatinib successfully inhibited the phosphorylation of FRS2, a downstream FGFR signalling molecule (Figure 1(b)). In addition, lenvatinib also inhibited the phosphorylation of ERK and p38 in HepG2 cells (Figure 1(b)). At the same time, there was no significant change in the phosphorylation of FRS2, ERK, and p38 in HepG2/R cells (Figure 1(b)). In order to identify the potential lncRNAs that may contribute resistance to lenvatinib, RNA sequencing was performed. Among various lncRNAs, lncXIST was found significantly upregulated in HepG2/R cells (Figure 1(c)). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis showed that several signalling pathways were highly enriched in HepG2/R cells (Figure 1(d)). Then, we measured the expression of lncXIST in various cells, and it was found that the expression of lncXIST in HCC cells was much higher than normal hepatocyte LO2 cells (Figure 1(e)). We also found that the IC50 value of lenvatinib in HCC cells is positively correlated with the expression of lncXIST (Figure 1(f)). To investigate the possible role of lncXIST in regulation HCC cell response to lenvatinib, siRNAs were applied to knockdown lncXIST. As indicated in Figure 1(g), the si-lncXIST#1 was the most efficient one, so it was used in the following experiments. We also forced the expression of lncXIST by transfecting HepG2 cells with a vector expressing lncXIST (Figure 1(h)). It was found that knockdown of lncXIST decreased HepG2/R cells IC50 to lenvatinib, while upregulation of lncXIST increased HepG2/R cells IC50 to lenvatinib (Figure 1(i)). Cell viability assays also confirmed that silencing of lncXIST markedly inhibited HepG2/R cell proliferation compared with control cells, and conversely, overexpression of lncXIST promoted proliferation of HepG2 cells with or without lenvatinib (Figure 1(j)). Taken together, those data suggest that lncXIST is upregulated in lenvatinib-resistant HepG2/R cells and contribute to resistance to lenvatinib.
Next, we examined whether lncXIST would affect cell apoptosis. Compared with the control cells, silencing of lncXIST promoted cell death of HepG2/R cells with or without lenvatinib treatment (Figure 2(a)). At the same time, forced expression of lncXIST inhibited the cell death of HepG2 cells caused by lenvatinib (Figure 2(a)). In order to confirm the type of cell death, caspase-3/9 activity assays were conducted. It was found that silencing of lncXIST promoted the activities of caspase-3/9, while overexpression of lncXIST inhibited the activities of caspase-3/9 under the treatment of lenvatinib (Figure 2(b)). We next examined whether lncXIST regulates the FGFR signalling. As shown in Figure 2(c), silencing of lncXIST inhibited the phosphorylation of FRS2, p38, and ERK in HepG2/R cells. Meanwhile, overexpression of lncXIST increased the phosphorylation of FRS2, p38, and ERK in HepG2 cells (Figure 2(c)). Thus, lncXIST affects HCC cell sensitivity to lenvatinib via regulation of apoptosis and FGFR signalling pathway.
In order to further investigate the mechanisms underlying the resistance to lenvatinib conferred by lncXIST, RNA sequencing was conducted. As indicated in Figures 3(a) and 3(b), various genes and signalling pathways were affected after knockdown of lncXIST in HepG2/R cells. Among them, NOD2 got our attention due to the highest-fold upregulation after knockdown of lncXIST which was further confirmed by RT-PCR (Figure 3(c)). Western blots also confirmed that knockdown of lncXIST led to upregulation of NOD2, while overexpression of lncXIST inhibited the NOD2 (Figure 3(d)). Additionally, we examined the distribution of lncXIST using subcellular fractionation analyses. qRT-PCR results showed that lncXIST is mainly distributed in the nucleus (Figure 3(e)), indicating that lncXIST may involve in the regulation of transcription. Amounting evidence suggests that lncRNAs can regulate the expression of genes via interaction with RNA binding proteins such as HuR, HMGB1, and EZH2 [10]. To investigate whether lncXIST could interact with those RNA-binding proteins, RNA immunoprecipitation (RIP) assays were conducted. It was revealed that lncXIST can bind with HuR, EZH2, and HMGB1. However, lncXIST interaction with EZH2 was stronger, suggesting that lncXIST interacted specifically with EZH2 in HepG2/R cells (Figure 3(f)). To further examine the correlation between lncXIST and EZH2, we measured the expression of EZH2 after silencing/overexpression of lncXIST. Surprisingly, knockdown or overexpression of lncXIST did not affect the EZH2 at both mRNA and protein levels (data not shown). Since EZH2 is a core subunit of polycomb repressive complex 2 (PRC2), it plays an essential role in regulating cancer cell response to drugs [11]. To study the role of EZH2 in HCC cell response to lenvatinib, the levels of EZH2 were examined. It was found that both the mRNA and protein levels of EZH2 were upregulated in HepG2/R cells compared to HepG2 cells (Figures 3(g) and 3(h)). Then, three siRNAs against EZH2 were applied to knockdown EZH2 in HepG2/R and Huh7 cells (Figure 3(i)). It was found that silencing of EZH2 increased the cell death induced by lenvatinib in both HepG2/R and Huh7 cells (Figure 3(j)). At the same time, cell viabilities showed that silencing of EZH2 decreased the cell viabilities of HepG2/R and Huh7 cells under the treatment of lenvatinib (Figure 3(k)). Next, we investigated the correlation between the EZH2 and NOD2. RT-PCR and western blots showed that downregulation of EZH2 led to the upregulation of NOD2 in HepG2/R and Huh7 cells (Figures 3(l) and 3(m)). Noteworthy, silencing of EZH2 also led to the inhibition of H3K27me3 (Figure 3(m)). Furthermore, chromatin immunoprecipitation (ChIP) was conducted, and it was found that knockdown of lncXIST attenuated the binding of EZH2 and H3K27 trimethylation levels across the promoter region of NOD2 (Figure 3(n)). These results suggest that lncXIST affects HCC cell response to lenvatinib, at least partly, via the epigenetic inhibition of NOD2 via interacting with EZH2 in HepG2/R cells.
Next, we investigated the role of NOD2 in regulating HCC cell response to lenvatinib. RT-PCR and western blots showed that both the mRNA and protein levels of NOD2 were lower in HepG2/R cells compared to HepG2 cells (Figures 4(a) and 4(b)). To further analyze the function of NOD2, we overexpressed the NOD2 in HepG2/R (Figures 4(c) and 4(d)). It was found that forced expression of NOD2 increased the cell death of HepG2/R cells under the treatment of lenvatinib (Figure 4(e)). Caspases' activities also confirmed that upregulation of NOD2 resulted in increased caspases activities in HepG2/R cells under the treatment of lenvatinib (Figure 4(f)). Moreover, cellular viability assay also indicated that forced expression of NOD2 decreased viabilities of HepG2/R cells under the treatment of lenvatinib compared to the control group (Figure 4(g)). Taken together, those data indicate that forced expression of NOD2 restored lenvatinib-resistant HepG2/R cell sensitivity to lenvatinib.
Next, we examined whether inhibition of NOD2 expression could affect the effects of knockdown of lncXIST on regulating HCC cell sensitivity to lenvatinib. As shown in Figures 5(a) and 5(b), although knockdown of lncXIST significantly increased NOD2 mRNA and protein levels, these effects were reversed by coexpression of two siRNAs against NOD2. It was found that knockdown of NOD2 reversed the effects of silencing of lncXIST on the cellular viabilities of HepG2/R cells under the treatment of lenvatinib (Figure 5(c)). Flow cytometry analysis also indicated that silencing of NOD2 reversed the effects of silencing of lncXIST on cell death induced by lenvatinib in HepG2/R cells (Figure 5(d)). Furthermore, caspases' activities also confirmed that the augmented caspase-3/9 activities caused by silencing of lncXIST under the treatment of lenvatinib could be reversed by knockdown of NOD2 in HepG2/R cells (Figure 5(e)). To further confirm the connective role of EZH2 between XIST and NOD2, we overexpressed EZH2 in HepG2 and Huh7 cells (Figure 5(f)). It was observed that overexpression of EZH2 led to the downregulation of NOD2 in both cells (Figures 5(f) and 5(g)). Overexpression of EZH2 also promoted the cellular viabilities of both HepG2 and Huh7 cells under the treatment of lenvatinib (Figure 5(h)). Moreover, overexpression of EZH2 also reduced the cellular death and activation of caspase-3/9 induced by lenvatinib in HepG2 and Huh7 cells (Figures 5(i) and 5(j)). Collectively, those findings indicate that lncXIST-EZH2-NOD2 axis confers resistance to lenvatinib in HCC cells.
To investigate the correlation between NOD2 and MAPK signalling, p38 inhibitor (SB202190) and ERK inhibitor (SCH772984) were applied. As shown in Figures 6(a) and 6(b), both inhibitors did not affect the downregulation of NOD2 caused by overexpression of lncXIST. Therefore, we hypothesized that NOD2 might act upstream of MAPKs. Interestingly, it was found that downregulation of NOD2 led to the activation of ERK but not p38 in HepG2 and Huh7 cells (Figure 6(c)). Meanwhile, overexpression of NOD2 reduced the levels of phosphor-ERK in HepG2 and Huh7 cells (Figure 6(d)). To confirm the role of ERK in regulating cell response to lenvatinib, SCH772984 was applied. Interestingly, SCH772984 treatment increased the cell death induced by lenvatinib in both HepG2 and Huh7 cells (Figure 6(e)). Thus, downregulation of NOD2 led to the activation of ERK that might be responsible for the resistance to lenvatinib in liver cancer cells.
HCC is one of the leading causes of cancer-related death. Lenvatinib is a multitargeted TKI that has been approved for the treatment of unresectable HCC. However, the overall response rate was only around 40% in HCC patients who received lenvatinib [12]. The clinical application of lenvatinib is often limited by drug resistance; therefore, it is necessary to unveil the mechanisms underlying the chemoresistance to lenvatinib. In the current study, we revealed that lncXIST confers resistance to lenvatinib in HCC cells. Mechanically, lncXIST can interact with EZH2 to repress the expression of NOD2. lncRNAs, a class of noncoding RNAs, widely exist in mammalian genomes and can be detected in the tissues, body fluids, and exosomes [13]. Amounting evidence suggests that lncRNAs play essential roles in various biological activities such as cell growth, differentiation, and cell death [14]. Various lncRNAs have also been identified as a regulator of HCC cell response to TKIs. For instance, upregulation of lncNIFK-AS1 conferred resistance to sorafenib in HCC cells [15]. LncMT1JP was able to promote lenvatinib resistance in HCC cells via inhibiting apoptosis [16]. lncXIST was identified as an oncogene in various cancers including the HCC. Liu and Xu reported that lncXIST promotes progression of HCC via sponging miR-200b-3p [17]. Dong et al. found that lncXIST can accelerate the growth of HCC cells via inhibiting miR-488 [18]. lncXIST has also been documented to affect cancer cell response to chemotherapy agents. Upregulation of lncXIST confers resistance to 5-FU and doxorubicin in colorectal cancer cells [19, 20]. In the present study, we revealed for the first time that lncXIST also confers resistance to lenvatinib in HCC cells. Previous studies have shown that lncRNAs regulate cancer cells sensitive to chemotherapeutics via various mechanisms. For instance, lncMT1JP promotes resistance to lenvatinib via acting as a competing endogenous RNA to miR-24-3p in HCC cells [16]. lncRNAs can also regulate gene transcription by recruiting histone modification enzymes or interacting with transcription factors. Chen et al. found that lncRNA CASC9 promoted resistance to gefitinib in NSCLC cells via epigenetic inhibition of DUSP1 [21]. In the current study, we revealed that lncXIST was able to bind with histone modification enzyme, EZH2, to inhibit the expression of NOD2. EZH2 is the core subunit of the PRC2 complex, which negatively regulate the gene expression via trimethylating of H3K27 [22]. It has been reported that EZH2 is overexpressed in HCC and is correlated with poor prognosis [23]. Previous studies found that EZH2 is able to regulate HCC cell sensitivity to sorafenib. For example, inhibition of EZH2 augmented the antitumor effects of sorafenib in HCC cells [24]. Interestingly, various studies found that EZH2 could interact with lncRNAs. Zhang et al. reported that lncUPK1A-AS1 was able to interact with EZH2 and promotes the proliferation of HCC cells [25]. In the current study, we found that knockdown of EZH2 could partially reversed lenvatinib resistance and promoted cell death of HCC cells. Our findings are in accordance with previous studies indicating that EZH2 could be used as a target to overcome lenvatinib resistance in HCC cells. We also conducted RNA sequencing to find the target genes of lncXIST, and NOD2 was identified. NOD2 is one of the pivotal innate immune sensors, which can recognize pathogen infection and induce subsequent innate immune response [26]. NOD2 acts as a tumor suppressor and was found to protect mice from inflammation and obesity-dependent HCC [27]. NOD2 has also been reported to inhibit tumorigenesis and increase chemosensitivity of HCC cells via targeting AMPK pathway [28]. In line with those studies, we also found that upregulation of NOD2 prompted cell death induced by lenvatinib in HCC cells. Therefore, the levels of NOD2 might be applied as a marker to predict the lenvatinib response in HCC. Noteworthy, another study found that NOD2 was upregulated and activated in HCC tissues, and high expression of NOD2 was correlated with poor prognosis in HCC patients [29]. This discrepancy reveals the complex role of NOD2 in regulating the tumorigenesis of HCC, and more investigations are required. We also revealed that inhibition of NOD2 led to the activation of ERK in liver cancer cells. Similar to our findings, ERK activation was found enhanced in NOD2-/- macrophages [30]. Moreover, ERK signalling was also found upregulated in NOD2-/- mice [31]. In addition, our data also indicated that activation of ERK was correlated with resistance to lenvatinib. This finding is in accordance with a previous study which also showed that ERK signalling conferred resistance to lenvatinib in liver carcinoma cells [32]. Hence, targeting ERK signalling might be a strategy to overcome lenvatinib resistance. Till now, there are many possible strategies to target the lncXIST/EZH2/NOD2/ERK axis. Many specific inhibitors against EZH2 have been developed and studied in the preclinical setting. For example, three EZH2 inhibitors, tazemetostat (EPZ-6438), GSK2816126, and CPI-1205, have moved into phase I/phase II clinical trials in patients with non-Hodgkin lymphoma and genetically defined solid tumors [33]. Clinical data showed that those EZH2 inhibitors are relatively safe. Noteworthy, another study found that EZH2 inhibitors prevented emergence of acquired resistance and augmented chemotherapeutic efficacy in both chemosensitive and chemoresistant models of small cell lung cancer [34]. In addition, SB 9200 is a novel, first-in-class oral modulator of innate immunity that is believed to act via the activation of the NOD2 pathways [35]. Although SB 9200 has a broad-spectrum antiviral activity, whether it possesses antitumor activities has not been reported yet. Amounting evidence suggests that angiogenesis and signalling through the ERK have been reported to play essential roles in hepatocarcinogenesis [36]. Sorafenib is an ERK inhibitor that has been approved for the treatment of liver carcinoma. Hence, it would be interesting to test whether those agents could enhance the efficacy of lenvatinib in liver carcinoma cells. There are some limitations of our study. Firstly, our studies are conducted using in vitro assays. It would be interesting to validate our findings in vivo. Secondly, there might be other genes that were affected by the lncXIST/EZH2 and affect HCC cell response to lenvatinib; it is worthy for further investigations.
In this study, we found that lncXIST promotes resistance to lenvatinib in HCC cells. Mechanically, lncXIST interacts with EZH2 to inhibit the expression of NOD2. Overexpression of NOD2 silencing could reverse the effects of inhibition of lncXIST on HCC cell sensitivity to lenvatinib. Our findings provide novel insights into the lncXIST/EZH2/NOD2/ERK axis in regulating HCC lenvatinib resistance (Figure 6(f)). | true | true | true |
||
PMC9596773 | Michael Claxton,Michela Pulix,Michelle K. Y. Seah,Ralph Bernardo,Peng Zhou,Sultan Aljuraysi,Triantafillos Liloglou,Philippe Arnaud,Gavin Kelsey,Daniel M. Messerschmidt,Antonius Plagge | Variable allelic expression of imprinted genes at the Peg13, Trappc9, Ago2 cluster in single neural cells 10.3389/fcell.2022.1022422 | 12-10-2022 | genomic imprinting,allelic expression,single-cell analysis,neurosphere,neural stem cell,Peg13,Trappc9,Ago2 | Genomic imprinting is an epigenetic process through which genes are expressed in a parent-of-origin specific manner resulting in mono-allelic or strongly biased expression of one allele. For some genes, imprinted expression may be tissue-specific and reliant on CTCF-influenced enhancer-promoter interactions. The Peg13 imprinting cluster is associated with neurodevelopmental disorders and comprises canonical imprinted genes, which are conserved between mouse and human, as well as brain-specific imprinted genes in mouse. The latter consist of Trappc9, Chrac1 and Ago2, which have a maternal allelic expression bias of ∼75% in brain. Findings of such allelic expression biases on the tissue level raise the question of how they are reflected in individual cells and whether there is variability and mosaicism in allelic expression between individual cells of the tissue. Here we show that Trappc9 and Ago2 are not imprinted in hippocampus-derived neural stem cells (neurospheres), while Peg13 retains its strong bias of paternal allele expression. Upon analysis of single neural stem cells and in vitro differentiated neurons, we find not uniform, but variable states of allelic expression, especially for Trappc9 and Ago2. These ranged from mono-allelic paternal to equal bi-allelic to mono-allelic maternal, including biased bi-allelic transcriptional states. Even Peg13 expression deviated from its expected paternal allele bias in a small number of cells. Although the cell populations consisted of a mosaic of cells with different allelic expression states, as a whole they reflected bulk tissue data. Furthermore, in an attempt to identify potential brain-specific regulatory elements across the Trappc9 locus, we demonstrate tissue-specific and general silencer activities, which might contribute to the regulation of its imprinted expression bias. | Variable allelic expression of imprinted genes at the Peg13, Trappc9, Ago2 cluster in single neural cells 10.3389/fcell.2022.1022422
Genomic imprinting is an epigenetic process through which genes are expressed in a parent-of-origin specific manner resulting in mono-allelic or strongly biased expression of one allele. For some genes, imprinted expression may be tissue-specific and reliant on CTCF-influenced enhancer-promoter interactions. The Peg13 imprinting cluster is associated with neurodevelopmental disorders and comprises canonical imprinted genes, which are conserved between mouse and human, as well as brain-specific imprinted genes in mouse. The latter consist of Trappc9, Chrac1 and Ago2, which have a maternal allelic expression bias of ∼75% in brain. Findings of such allelic expression biases on the tissue level raise the question of how they are reflected in individual cells and whether there is variability and mosaicism in allelic expression between individual cells of the tissue. Here we show that Trappc9 and Ago2 are not imprinted in hippocampus-derived neural stem cells (neurospheres), while Peg13 retains its strong bias of paternal allele expression. Upon analysis of single neural stem cells and in vitro differentiated neurons, we find not uniform, but variable states of allelic expression, especially for Trappc9 and Ago2. These ranged from mono-allelic paternal to equal bi-allelic to mono-allelic maternal, including biased bi-allelic transcriptional states. Even Peg13 expression deviated from its expected paternal allele bias in a small number of cells. Although the cell populations consisted of a mosaic of cells with different allelic expression states, as a whole they reflected bulk tissue data. Furthermore, in an attempt to identify potential brain-specific regulatory elements across the Trappc9 locus, we demonstrate tissue-specific and general silencer activities, which might contribute to the regulation of its imprinted expression bias.
Genomic imprinting has long been recognized as a paradigm of epigenetic regulation of a specific subset of ∼200 genes that have important roles in mammalian embryogenesis, regulation of nutrient supply and demand between mother and offspring, as well as brain development and neural functions (Ferguson-Smith, 2011; Peters, 2014; Tucci et al., 2019). Imprinted genes are defined by their parent-of-origin dependent mono-allelic or strongly biased allelic (>70%) expression. This expression bias towards a specific parental allele is a consequence of DNA methylation marks that are differentially established in either the male or female germ cells, respectively, at a defined set of CpG-rich islands (CGIs). Such germline differentially methylated regions (gDMRs) are maintained after fertilization in the somatic cells of the developing and adult offspring; they are only erased in its developing germline cells for re-setting according to its sex and transmission to the next generation (Plasschaert and Bartolomei, 2014). The germline DMRs often regulate the allelic expression of a cluster of neighboring genes and are, therefore, also called imprinting control regions (ICRs) (Ferguson-Smith, 2011; Peters, 2014; Tucci et al., 2019). In addition to DNA methylation, the mechanisms involved in the regulation of imprinted gene expression comprise histone modifications, non-coding RNAs and boundary or insulator elements that are recognized by CTCF, a methylation-sensitive DNA binding factor. CTCF binding at unmethylated sites within DMRs of imprinted genes has been shown to regulate access to tissue-specific enhancers and the formation of allelic topologically associated domains (TADs), thereby controlling the expression of neighboring genes in an allele-specific way (Bell and Felsenfeld, 2000; Lleres et al., 2019). The imprinting status of most genes is conserved between human and mouse, although some genes do not show an allelic expression bias in one or the other species (Court et al., 2014; Tucci et al., 2019). Furthermore, while many imprinted genes have the same strong parental allele-specific expression bias in all tissues analyzed, others show tissue-specific imprinting effects, examples of which are Gnas in defined brain regions, endocrine glands and proximal renal tubules, Ube3a in neurons (but not glia and peripheral tissues), as well as Ago2, Trappc9 and Chrac1 in brain (Yamasaki et al., 2003; Weinstein et al., 2010; Mabb et al., 2011; Peters, 2014; Perez et al., 2015). The imprinting status of some genes can also be changed in specific cell types. Dlk1 is a paternally expressed gene with important functions in adipose tissue development, metabolic regulation and neurogenesis (Peters, 2014). However, its imprinting status is lost in neural stem cells (NSCs) of the subventricular zone and hippocampal dentate gyrus with bi-allelic expression of Dlk1 being required for postnatal and adult neurogenesis in these stem cell niches (Ferron et al., 2011; Montalban-Loro et al., 2021). The mechanism for this change in allelic expression status involves postnatal gain of methylation at the gDMR/ICR of the locus (Ferron et al., 2011). The assessment of allelic expression biases of imprinted genes requires the generation of hybrid mice in crosses between different strains that carry a sufficient number of single nucleotide polymorphisms (SNPs) to be able to identify the parental alleles. Until recently, it was only possible to investigate the allelic expression of genes on the level of bulk primary cell culture or tissue lysates, which often contain different cell types. Although many imprinted genes show an almost exclusive parental bias of ∼90% and are considered mono-allelic in their expression, recent RNA-seq studies have revealed a number of genes with a weaker parental bias of ∼70%, which was used as a threshold for inclusion in the ‘imprinted gene’ category (Babak et al., 2015; Bonthuis et al., 2015; Crowley et al., 2015; Perez et al., 2015; Bouschet et al., 2016; Andergassen et al., 2017; Huang et al., 2017). These data raise questions about how such bulk tissue-level expression biases are reflected on a single cell level, and about the biological significance of such findings (Perez et al., 2016). Several single-cell expression scenarios can potentially lead to a parental allele-specific expression bias of ∼70% in tissues (Bonthuis et al., 2015; Perez et al., 2016), including 1) all cells show the same biased bi-allelic expression, 2) the tissue consists of a mixture of cells with mono-allelic and equal bi-allelic expression, 3) the tissue consists of an unbalanced mixture of cells with respectively mono-allelic paternal and mono-allelic maternal expression, which can be due to differential promoter usage as is the case for Grb10 in neurons versus glial cells and peripheral tissues (Yamasaki-Ishizaki et al., 2007; Sanz et al., 2008; Garfield et al., 2011). To address this question, two types of approaches have been applied recently. Using SNP-FISH in situ hybridization, which employs SNP-specific oligonucleotides, Ginart et al. (2016) were able to distinguish mono-versus bi-allelic expression of H19 and Igf2 in fixed fibroblasts and heart tissue via imaging. With a different in situ hybridization technique, using intronic RNAscope probes for nascent RNA in nuclei, Bonthuis et al. (2015) showed a mixture of cells with mono-allelic and bi-allelic expression of specific imprinted genes in brain sections. Novel single-cell RNA-seq methods have also been applied to analyze imprinted gene expression in single cortical neurons after labeling them with fluorescent proteins and FACS sorting (Laukoter et al., 2020). Their findings indicate some degree of variability of allelic expression in individual neurons, depending on the imprinted gene analyzed. While Meg3 and Snrpn showed the expected mono-allelic expression in almost all neurons, Inpp5f and Impact were mono-allelic paternally expressed in the majority of cells with smaller numbers of neurons displaying bi-allelic or even mono-allelic maternal expression (Laukoter et al., 2020). Thus, these initial studies indicate that the imprinted expression status of a gene, as determined on a tissue level, might not be reflected in all its cells. The Peg13-Kcnk9-Trappc9 imprinting cluster (schematically shown in Figure 6A) on mouse chromosome 15/human chromosome 8 consists of several genes with neurodevelopmental functions; mutations of these cause disorders in both species. At the center of the locus, the non-coding RNA Peg13 is expressed from the paternal allele, starting at an unmethylated CGI promoter and DMR located within an intron of Trappc9 (Smith et al., 2003; Ruf et al., 2007). Germline-derived methylation silences Peg13 on the maternal allele. Complete deletion of Peg13 on the unmethylated paternal allele is lethal in mice, while the same mutation on the maternal allele does not cause a phenotype (Keshavarz and Tautz, 2021). A milder behavioral phenotype is observed with a mutation that truncates the non-coding RNA (Keshavarz and Tautz, 2021). The second gene of the cluster, Kcnk9, for which imprinting is conserved between human and mouse, encodes the two-pore domain potassium channel subunit Task3 (Court et al., 2014). Kcnk9 is expressed with a strong maternal allelic bias in adult brain (Ruf et al., 2007; Court et al., 2014; Cooper et al., 2020). Mutations of the gene cause Birk-Barel intellectual disability syndrome in humans and behavioral abnormalities in mice (Linden et al., 2007; Barel et al., 2008; Cooper et al., 2020). The three genes Trappc9, Chrac1 and Ago2 show an imprinted bias of expression from the maternal allele in mouse brain, but are not imprinted in human (Court et al., 2014; Babak et al., 2015; Bonthuis et al., 2015; Crowley et al., 2015; Perez et al., 2015; Bouschet et al., 2016; Andergassen et al., 2017; Huang et al., 2017). The mechanisms underlying the predominantly maternal expression of these three genes in mouse brain are currently unclear. Trappc9 encodes a subunit of the intracellular trafficking protein particle II complex (TrappII), mutations of which lead to a neurodevelopmental disorder in humans and mice, which includes symptoms of postnatal microcephaly, intellectual disability and speech impairment (Ke et al., 2020; Liang et al., 2020; Wilton et al., 2020; Aslanger et al., 2022). Ago2 forms a subunit of the RNA-induced silencing complex (RISC). Heterozygous mutations in humans result in a range of neurological phenotypes while homozygous mutation in mice is embryonic lethal (Liu et al., 2004; Lessel et al., 2020). Little is known about the chromatin accessibility factor Chrac1. In this study, we analyzed the allelic expression of these genes in bulk tissue and neurosphere lysates comparatively to single NSCs and differentiated neurons. We especially focused on Peg13, Trappc9 and Ago2 as examples of strongly and moderately biased imprinted genes. We found variability of allelic expression in individual cells, which was more pronounced for Trappc9 and Ago2 than for Peg13. All categories of expression from mono-allelic maternal to equal bi-allelic to mono-allelic paternal, as well as biased bi-allelic states, were identified for Trappc9 and Ago2 in single cells. For Peg13, a majority of cells showed the expected mono-allelic paternal or paternally biased bi-allelic expression, but a small number of cells deviated from this status, displaying equal bi-allelic or even a maternally biased expression. However, considering the whole population of single cells analyzed, we find that the imprinted gene expression status approximates the findings from bulk tissue lysates. Additionally, we determined the transcriptional start site of Trappc9 and investigated potential transcript variants as well as regulatory regions located within the locus, which led to the identification of sequence elements with a silencing function in primary neurons and/or fibroblasts.
Mouse strains C57BL/6J and Cast/EiJ were bred and maintained in the Babraham Institute Biological Support Unit. Ambient temperature was ∼19–21°C and relative humidity 52%. Lighting was provided on a 12 h light: 12 h dark cycle including 15 min “dawn” and “dusk” periods of subdued lighting. After weaning, mice were transferred to individually ventilated cages with 1–5 mice per cage. Mice were fed CRM (P) VP diet (Special Diet Services) ad libitum and received seeds (e.g., sunflower, millet) at the time of cage-cleaning as part of their environmental enrichment. Breeding and maintenance of these strains were performed under licenses issued by the Home Office (United Kingdom) in accordance with the Animals (Scientific Procedures) Act 1986 and were approved by the Animal Welfare and Ethical Review Body at the Babraham Institute. Tissues were collected from newborn or adult mice and either frozen for molecular biology or processed for cell culture. Frozen tissues from C57BL/6J and Mus musculus molossinus JF1 hybrid mice were kindly provided by Dr Philippe Arnaud, Université Clermont Auvergne, France.
Neurosphere culture was performed as described previously (Ferron et al., 2007; Chojnacki and Weiss, 2008) with slight modifications. Briefly, hippocampi were dissected from newborn mouse brain in ice-cold neurosphere growth medium (DMEM/F12 (Gibco) supplemented with 0.6% w/v glucose, 0.1% NaHCO3, 5 mM HEPES, 2 mM L-GIn, 100 U/ml penicillin, 0.1 mg/ml streptomycin, 1x B27 (Gibco), 10 ng/ml FGF-2 (Peprotech), 20 ng/ml EGF (Peprotech), 4 mg/ml BSA (Sigma) and then transferred into Accutase (Gibco) for dissociation into a single-cell suspension by gentle trituration. Following centrifugation at 200 g for 5 min, cells were resuspended in growth medium and plated at a density of 3,000 cells/cm2 in suspension cell culture dishes (Corning). Neurospheres were allowed to grow for 6–8 days with intermittent medium supplementation before passaging via Accutase dissociation. Cells were re-plated at a lower density of 1,500–2,000 cells/cm2. Neurospheres could be stored in liquid nitrogen after freezing in growth medium with 10% DMSO. For bulk or single-cell gene expression analysis, neurospheres at early passage numbers (P3–P5) were used. Neurospheres were differentiated into neurons at the point of passaging by seeding a single cell suspension on Poly-L-Lysine (Sigma) coated dishes in differentiation medium (growth medium without EGF, FGF-2 and BSA, but containing 1% FBS). For single-neuron analysis, selection against replicating glial cells was started after 2 days of culture with 2 μM Cytosine β-D-arabinofuranoside (AraC) (Sigma) as described in the next paragraph. Primary hippocampal neurons were cultured as described (Beaudoin et al., 2012; Ioannou et al., 2019) with modifications. Hippocampi were dissected from newborn mouse brain in ice-cold dissection medium [HBSS (Sigma) supplemented with 0.1% w/v glucose, 10 mM Hepes pH 7.4, 1% Na-pyruvate] and the tissue dissociated by adding an equal volume of 2x Papain stock solution (Worthington) at 37°C for 20 min. The supernatant was removed carefully, the tissue gently washed with plating medium [MEM (Gibco) supplemented with 0.45% glucose, 10% FBS, 1% Na-pyruvate, 2 mM Glutamine, 100 U/ml penicillin, 0.1 mg/ml streptomycin] and then carefully triturated in fresh plating medium. The dissociated tissue was rinsed through a 70-μm cell strainer (Corning) and the collected cells centrifuged at 200 g for 5 min. Cells were resuspended in neuronal medium [Neurobasal medium (Gibco) supplemented with 2 mM glutamine, 100 U/ml penicillin, 0.1 mg/ml streptomycin, 1x B27 (Gibco)] and plated in Poly-L-Lysine (Sigma) coated dishes at a density of 60,000 cells/cm2. Medium was replaced the following day, and on day two selection against replicating non-neuronal cells was started with neuronal medium containing 2 μM AraC. Half the medium was replaced with fresh neuronal medium every other day to dilute out the AraC. Mouse embryonic fibroblasts were prepared as described (Matise et al., 2000) and cultured from frozen stocks in Hepes-buffered DMEM (Sigma) supplemented with 10% FBS, 2 mM Glutamine, 100 U/ml penicillin, 0.1 mg/ml streptomycin.
The promoter-reporter gene plasmids were transfected into fibroblasts and into primary hippocampal neurons after 7 days of culture using Lipofectamine 2000 (Invitrogen). The firefly luciferase-based test constructs were mixed with a Renilla luciferase control plasmid (pGL4.74, Promega) at a 100:1 ratio to normalize for transfection efficiency. Cells were lysed 48 h after transfection and luciferase activities were measured using the Dual-Luciferase® Reporter Assay System (Promega) on a Glomax Multi Detection System (Promega).
RNA was isolated from neurospheres and tissues using TRIzol™ reagent (Invitrogen) or RNeasy Plus Mini kit (Qiagen). Samples were treated with DNAse I to remove any traces of DNA before cDNA was synthesized with ProtoScript® II Reverse Transcriptase (New England Biolabs) or SuperScript III™ Reverse Transcriptase (Invitrogen) using random hexamer primers, if not otherwise stated. PCR was performed using GoTaq® Hot Start Polymerase (Promega) or Q5™ High-Fidelity DNA Polymerase (New England Biolabs). For the bulk neurosphere and tissue gene expression analysis we used 1 μg of total RNA in reverse transcription reactions, which were then diluted 5-fold for endpoint PCR (initial denaturation 98°C, 30 s; 30 cycles of 98°C, 50–72°C annealing (primer dependent); 72°C, 30 s; final extension 72°C, 2 min) to obtain enough products for pyrosequencing analysis. 5′-RACE experiments were undertaken with ExactSTART™ Eukaryotic mRNA 5′&3′ RACE Kit (Epicentre/Illumina) on adult mouse brain RNA. The 5′-RACE cDNA was then amplified with a Trappc9-specific reverse primer (Pr_05RV; Supplementary Table S1) and a kit-supplied RACE 5′-linker primer, followed by cloning of PCR products into TOPO®-plasmids (Invitrogen) and sequencing.
SNPs in cDNA from tissues and neurospheres of hybrid mice, as well as genomic DNA for methylation analysis after bisulfite treatment, were sequenced using a PyroMark® Q96 ID instrument (Qiagen). PCR and sequencing primers (Supplementary Table S1) were designed using PyroMark Assay Design Software 2.0. Biotinylated PCR products were immobilized on streptavidin-coated beads for cleanup and sequenced using PyroMark® Gold Q96 reagents (Qiagen) following the manufacturer’s protocols.
Single C57BL/6J × Cast/EiJ neurosphere cells were obtained via dissociation with Accutase, dilution in growth medium and either FACS sorting or manual isolation via capillary action under a microscope using a protocol originally developed for oocytes (Lorthongpanich et al., 2013; Cheow et al., 2015; Cheow et al., 2016). Single neurons from differentiated neurospheres were collected after 7 days of culture in differentiation medium, which included 5 days of AraC treatment. Neurons were dissociated from culture dishes using Trypsin/EDTA (Sigma), diluted in differentiation medium and single cells isolated manually via capillary action under a microscope. Single cells were transferred into PCR tubes containing 5 μl of lysis buffer [CellsDirect Resuspension and Lysis Buffer, 10:1 (Invitrogen)] and incubated at 75°C for 10 min (Lorthongpanich et al., 2013; Cheow et al., 2015; Cheow et al., 2016). cDNA was synthesized at 37°C by adding an equal volume of a 2x reverse transcription master mix using MultiScribe™ Reverse Transcriptase (Invitrogen) and hexamer primers. This was followed by protease treatment (Qiagen Protease) as described (Lorthongpanich et al., 2013; Cheow et al., 2015; Cheow et al., 2016) to remove chromatin-associated proteins from genomic DNA. Next, we perform single-cell restriction analysis of methylation (SCRAM), using BstUI, to digest unmethylated CpG sites of genomic DNA. This provides the additional option of analyzing DNA methylation of CGIs in the single cells (Lorthongpanich et al., 2013; Cheow et al., 2015; Cheow et al., 2016). A final Proteinase K digest was carried out before undertaking multiplex PCR pre-amplification using a pool of primers for all target genes with the following conditions: initial denaturation 95°C, 10 min; 30 cycles of 95°C, 30 s; 60°C annealing/extension, 4 min. After pre-amplification, unincorporated primers were removed from the reactions by adding exonuclease I (Lorthongpanich et al., 2013; Cheow et al., 2015; Cheow et al., 2016). Pre-amplification reactions were then diluted 10-fold and aliquots were used to amplify individual target genes via nested-primer qPCR using 3 μl of diluted template and PowerUp SYBR Green 2x Master mix (Invitrogen) under the following conditions: initial denaturation 95°C, 10 min; 30 cycles of 95°C, 15 s; 60°C annealing/extension 1 min; melt curve analysis 60–95°C ramp 5 s/degree. Primers for target genes are listed in Supplementary Table S1. Where applicable, these qPCR products were purified using MinElute PCR® purification kit (Qiagen) and Sanger-sequenced for cDNA SNP expression analysis.
Genomic DNA was isolated from tissues through Proteinase K (100 μg/ml) digest in lysis buffer (100 mM Tris, 5 mM EDTA, 200 mM NaCl, 0.2% SDS, pH 8.5) at 55°C overnight, followed by Phenol/Chloroform extraction, Ethanol precipitation and resuspension in TE buffer. Neurosphere DNA was obtained using TRIzol™ reagent (Invitrogen) in a follow-on step after initial RNA isolation via Phenol/Ethanol extraction, precipitation and resuspension in TE. For DNA methylation analysis of CpG sites, bisulfite conversion of unmethylated cytosines was carried out using the EZ DNA Methylation-Gold™ kit (Zymo Research) according to manufacturer instructions. The bisulfite-treated and purified DNA was then used for PCR amplification of CGI fragments, followed either by methylation analysis via direct pyrosequencing or cloning of PCR products into TOPO®-vectors (Invitrogen) and Sanger sequencing of individually cloned plasmid samples. Sanger sequencing results were further analyzed using the free online tool “QUantification tool for Methylation Analysis” (QUMA; http://quma.cdb.riken.jp/).
Potential brain-specific gene regulatory elements were identified from histone modification and CTCF ChIP assay data, as well as DNAse I and ATAC-seq hypersensitivity data for the newborn (P0) mouse brain in comparison to peripheral tissues. These data were extracted from the databases ENCODE3 (https://www.encodeproject.org/), UCSC Genome Browser (https://genome-euro.ucsc.edu/) and ENSEMBL (https://www.ensembl.org/Mus_musculus/Info/Index). Seven candidate brain-specific regulatory elements across the Trappc9-Peg13 locus were found; their genomic positions (mouse GRCm38/mm10 genome version) and features are listed in Supplementary Figure S6. Similarly, the Trappc9 promoter characteristics at exon 1 were noted. To test the functionality of these regulatory elements in transfected cells, promoter-reporter gene plasmids were generated. The regulatory elements were amplified from C57BL/6J genomic DNA using Q5™ High-Fidelity DNA Polymerase (New England Biolabs), cloned into TOPO®-plasmids (Invitrogen) and sequenced for confirmation. Similarly, four Trappc9 promoter fragments of different lengths were cloned (positions 73,061,805–73,060,204 bp, 73,061,805–73,060,975 bp, 73,061,418–73,060,204 bp and 73,061,418–73,060,975 bp in GRC38/mm10). The firefly luciferase-encoding pGL4.23 [luc2/minP] vector (Promega) was used to generate reporter-gene constructs. First, the endogenous minimal promoter was removed via HindIII and NcoI digest and then replaced with a Trappc9 promoter fragment. The four Trappc9 promoter plasmids were tested in a preliminary reporter gene assay for their activity. The promoter fragment 73,061,418–73,060,975 bp, which avoids an upstream dinucleotide repeat sequence stretch and ends before the exon 1 splice donor site, showed the highest activity and was used in further experiments in combination with the identified regulatory elements. The regulatory elements were cloned into the pGL vector at the BamHI site downstream of the Luc2 reporter gene to reflect the same relative orientation to the Trappc9 promoter as in the genome. Only the Reg-E element was cloned upstream of the Trappc9 promoter as shown in Figure 6B.
The data for the promoter-reporter gene assays were analyzed using GraphPad Prism v.9.3 software. Data were analyzed for outliers using the ROUT method. The datasets were then analyzed for normality using the Shapiro-Wilkinson test. For non-parametric datasets, Mann-Whitney U-tests were performed in comparisons of the Reg-element datasets to the basic Trappc9 promoter dataset. For parametric datasets, unpaired t-tests were performed.
The core imprinted gene of this cluster on mouse chromosome 15, Peg13, is located within intron 17 of Trappc9 and is transcribed into a non-coding RNA (Smith et al., 2003; Wang et al., 2008). For Trappc9 itself, several alternatively spliced and truncated transcript variants have been described and annotated on ENSEMBL, including alternative first exons (Figure 1A). Furthermore, while some transcript variants were found to be predominantly expressed from the maternal allele in brain tissue, the truncated variant 203, which ends shortly after Peg13 in intron 17, was identified as a paternal allele-specific transcript in RNA-seq studies (Gregg et al., 2010; Hsu et al., 2018). We set out to confirm these variants using RT-PCR across specific Trappc9 exons. We readily detected the full-length transcript 202 in brain and other tissues of newborn mice but were not able to confirm the alternative first exon of the 201 transcript (Figure 1B). Exon 2 of the 202 transcript contains a conserved translational start codon across mammalian species. To further investigate potential alternative transcriptional start sites, we undertook 5′-RACE PCR and sequencing, which revealed several transcriptional start sites within the exon 1 5′-UTR of variant 202 (Figure 1C), but we did not detect the alternative exon 1 of variant 201. Sequencing of PCR products also showed alternative splicing of exon 5, which was missing in some kidney and spleen cDNAs, in line with ENSEMBL annotations. Additionally, we investigated the alternative, truncated Trappc9 splice forms 206 and 203, which terminate in intron 17 upstream and downstream of Peg13, respectively (Figure 1A, Supplementary Figure S1). We were unable to detect any cDNA containing the shared exon 17 in combination with 3′-UTR exons of variants 206 and 203 in tissues (Supplementary Figure S1), or containing shared exon 16 in single neurosphere cells. In conclusion, we have experimentally verified the transcriptional start sites of Trappc9 located in exon 1 of the 202 variant as well as alternative splicing of exon 5 but found no evidence of an alternative promoter or truncated transcript variants.
Apart from Peg13, which constitutes a canonical imprinted gene with mono-allelic paternal expression in a wide range of tissues, the other genes of the cluster have been characterized via RNA-seq as tissue-specifically imprinted with biased expression from the maternal allele in mouse brain (Babak et al., 2015; Perez et al., 2015; Andergassen et al., 2017). We set out to validate these findings through SNP pyrosequencing and additionally included samples of primary NSC (neurosphere) cultures from newborn mice (Supplementary Figures S2A,B), since some imprinted genes have been shown to become bi-allelically expressed specifically in postnatal and adult NSCs (Ferron et al., 2011; Montalban-Loro et al., 2021). We used tissues and cultured hippocampal neurospheres from reciprocal crosses of C57BL/6J and Mus musculus castaneus (Cast/EiJ) newborn F1 hybrids. While Peg13 showed the expected, almost exclusive (80%–90%) expression from the paternal allele in brain, kidney and neurospheres, the other genes of this cluster displayed tissue-specific imprinted expression (Figure 2). Trappc9 and Ago2 were predominantly (70%–80%) transcribed from the maternal allele in brain, but showed equal bi-allelic expression in kidney, while Kcnk9 was expressed almost exclusively (>90%) from the maternal allele in both tissues. Due to unavailability of SNPs between C57BL/6J and Cast/EiJ strains, we analyzed Chrac1 in reciprocal crosses of C57BL/6J and Mus musculus molossinus (JF1) F1 tissue samples and found it to be bi-allelically expressed with only a small bias (<70%) towards the maternal allele in newborn brain (Supplementary Figure S3). We also confirmed brain-specific imprinted expression of Trappc9 in these hybrids (Supplementary Figure S3). Unexpectedly and in contrast to brain tissue as a whole, Ago2 and Trappc9 were not imprinted in the NSC cultures, but showed equal bi-allelic or only slightly biased (<70%) expression (Figure 2), which is reminiscent of Dlk1 (Ferron et al., 2011; Montalban-Loro et al., 2021). Kcnk9 allelic expression in neurospheres was inconclusive and prone to strain-specific biases (Figure 2). We undertook pyrosequencing and/or Sanger sequencing of bisulfite-treated DNA to address DNA methylation states at the CGIs of the genes in neurospheres. The germline differentially methylated region (DMR) at Peg13 was maintained in NSCs with methylation observed on the maternal allele (Supplementary Figure S4A), in line with brain tissue observations (Ruf et al., 2007; Xie et al., 2012). The Trappc9 CGI1, located at the promoter/exon1, was unmethylated in NSCs, brain and kidney (Supplementary Figure S4B), while the CGI2 at exon 2 was fully methylated on both alleles in NSCs and brain (Supplementary Figure S4C). The promoter CGIs at Ago2, Chrac1 and Kcnk9 were also unmethylated in NSCs (Supplementary Figures S5A−C), which is in line with human brain data (Court et al., 2014) and their status as actively transcribed genes. This excludes any secondary DMRs at this imprinted gene cluster. Overall, our data confirm brain-specific imprinting, i.e., preferential expression from the maternal allele, of Trappc9 and Ago2 in mouse, while maternal allelic expression of Kcnk9 occurs in brain and some peripheral tissues, e.g., kidney. Unexpectedly, Trappc9 and Ago2 have no allelic expression bias in hippocampal neurosphere cultures.
Analysis of imprinted gene expression on a bulk tissue level raises the question of whether the observed allelic bias is reflected in every cell of the lysate in the same way, or whether individual cells differ in their mono-/bi-allelic transcriptional status of the gene and, thus, deviate from the tissue average. Large-scale single-cell imprinted gene expression analysis is still in its infancy, but novel approaches indicate that not all cells of a tissue might show the same allelic expression status (Martini et al., 2022). To address this question specifically for the Peg13 imprinting cluster, we isolated single NSCs from C57BL/6J × Cast/EiJ neurospheres as well as single neurons differentiated from these in vitro. We then undertook qRT-PCR using the sc-GEM (single-cell analysis of genotype, expression and methylation) technique (Cheow et al., 2016) and Sanger sequencing to determine allelic SNP expression for Peg13, Trappc9 and Ago2. As a further confirmation of the NSC phenotype, we also verified marker gene expression (Hochgerner et al., 2018) in single neurosphere cells (Supplementary Figure S2C). Analyzing ∼50 NSCs and neurons, we found a variety of mono-allelic and bi-allelic expression states in individual cells. For Peg13, 39% of the NSCs showed mono-allelic paternal expression and 40% of the cells had transcripts predominantly from the paternal allele, but additionally a small amount of maternal transcripts (Figures 3A–C). However, we also detected small numbers of cells with equal bi-allelic, predominantly maternal, or even mono-allelic maternal expression of Peg13 (Figures 3B,C), which indicates a surprising heterogeneity of imprinted expression between individual NSCs. Proportionately, the 79% of cells with mono-allelic and biased paternal expression approximate the bulk neurosphere expression bias of 87% paternal transcripts (Figure 2). In in vitro differentiated neurons, the cellular heterogeneity was reduced as almost all cells showed mono-allelic or paternal bias of Peg13 expression (Figure 3B), much in line with the paternal bias of 89% in brain tissue as a whole (Figure 2). For Trappc9 the allelic expression varied considerably in single NSCs. We observed mono-allelic maternal or paternal Trappc9 expression in 10% of the cells respectively, while 29% displayed equal bi-allelic transcription (Figures 4A–C). Around half of the NSCs showed biased bi-allelic expression: 21% predominantly maternal, 31% predominantly paternal (Figure 4B). Considering all single-NSC categories in proportion, the data are in line with the bulk neurosphere analysis (Figure 2) as no clear overall parental allele bias was observed in either dataset. In differentiated neurons, the proportion of cells with equal bi-allelic expression was reduced to 11%, while another 11% showed mono-allelic maternal and 17% mono-allelic paternal Trappc9 expression (Figure 4B). Taking into account the proportions of neurons with biased bi-allelic expression (33% maternal, 28% paternal bias), the single-neuron dataset depicts an overall equal bi-allelic Trappc9 expression, which is in contrast to the maternal expression bias of 78% that was observed in whole-brain lysate (Figure 2). This discrepancy might be due to the brain lysate containing additional cell types (e.g., astrocytes, microglia, oligodendrocytes) and a wider range of neuron types from different brain sub-regions as compared to the in vitro differentiated neurons, which were derived from hippocampal neurospheres. Ago2 expression in single NSCs was similarly variable as Trappc9 expression. 10% of the NSCs displayed mono-allelic maternal, 16% mono-allelic paternal and 12% equal bi-allelic expression (Figures 5A–C). Most NSCs had a biased bi-allelic expression of Ago2 (29% predominantly maternal, 33% predominantly paternal). Overall, the proportions of NSCs falling into the various expression categories did not indicate a clear allele bias and are in line with the equal bi-allelic Ago2 expression found in bulk neurosphere samples (Figure 2). Compared to the NSCs, among the differentiated neurons more cells showed equal bi-allelic (21%), mono-allelic maternal (16%) and biased maternal (32%) expression (Figure 5B), while the proportions of neurons with mono-allelic paternal (9%) and biased paternal (23%) expression were reduced (Figure 5B). However, as with Trappc9, the in vitro differentiated neuron categories did not overall reflect the same strong bias of 75% maternal allele-specific Ago2 transcripts that were detected in whole-brain tissue (Figure 2). Since the non-coding RNA Peg13 is transcribed from the core imprinting regulatory region (DMR) of the locus, findings of Peg13 expression other than mono-allelic paternal or paternally biased bi-allelic (Figures 3B,C) were unexpected. This raises the question of whether there is a specific pattern of allelic expression of the other imprinted genes of the locus associated with Peg13 transcription from the maternal allele. When analyzing the expression status of Trappc9 and Ago2 in those cells that showed equal bi-allelic, maternally biased or mono-allelic maternal expression of Peg13, we did not find any specific patterns or correlations (Supplementary Table S2). Some of these cells displayed the expected maternal bias of Trappc9 and/or Ago2, but other allelic biases, including mono-allelic expression states, were also observed and in varying combinations within individual cells. In summary, our single-cell analysis of the three imprinted genes indicates a surprising variability of allelic expression states in individual cells, ranging from mono-allelic maternal to mono-allelic paternal transcription, even for the core imprinted gene of the locus, Peg13. Overall, Peg13 transcriptional states in the NSC and neuron populations matched the brain tissue level of allelic expression (∼89% paternal) very well as most cells displayed mono-allelic or strong paternal bias. Trappc9 and Ago2, which represent tissue-specifically imprinted genes with a maternal expression bias of ∼75% in brain, showed much more variability in their allelic transcriptional states in individual NSCs and neurons; all categories of allelic transcription were represented by substantial numbers of cells. Thus, our data do not support a model, in which a tissue-level imprinted gene expression status is reflected in each cell of the tissue in the same way. These findings might hint at a certain level of transcriptional noise or transient/random bursts of transcription, which might still be able to occur at alleles that are “silenced” by genomic imprinting (Varrault et al., 2020).
It is currently unclear how tissue-specific imprinting and maternal allele-biased expression of Trappc9 and Ago2 are regulated in the mouse brain. Furthermore, imprinting of the two genes is not conserved in humans (Court et al., 2014) and only homozygous mutations of TRAPPC9 cause a neurodevelopmental disorder, which is characterized by intellectual disability, speech impairment and microcephaly (Wilton et al., 2020; Aslanger et al., 2022). A potential mechanism could involve chromatin boundaries and CTCF-regulated access to tissue-specific enhancers as has been shown for the imprinted Igf2-H19 locus (Bell and Felsenfeld, 2000). Indeed, CTCF binding on the unmethylated paternal allele of the Peg13 DMR has been demonstrated in mouse brain and fibroblasts (Singh et al., 2011; Prickett et al., 2013) as well as in human brain (Court et al., 2014). In humans, CTCF regulates access of KCNK9 and PEG13 promoters to a brain-specific enhancer, most likely in a differential, allele-specific way (Court et al., 2014). However, since the imprinting status of the genes upstream of Peg13, i.e., Trappc9, Ago2 and Chrac1, differs between mouse and human brain, their regulation presumably involves enhancers that are not conserved or additional mechanisms. We, therefore, screened Encode3 mouse genome data for chromatin modifications and accessibility (Consortium et al., 2020; Gorkin et al., 2020) across the Trappc9-Peg13 locus in the UCSC Genome Browser and identified seven potential brain-specific regulatory elements with appropriate histone modification (H3K4 methylation, H3K27 acetylation, H3K9 acetylation), ATAC and DNAse I hypersensitivity marks (Supplementary Figure S6). To test these candidate elements in promoter-reporter gene assay, we first constructed a Luciferase plasmid that contained a 444 bp Trappc9 promoter fragment upstream of and including non-coding exon 1 (Figure 1C). The size of this promoter fragment is in line with standards from high-throughput testing of promoter-enhancer interactions in the mouse genome (Martinez-Ara et al., 2022). We positioned the candidate regulatory elements downstream or upstream of the promoter-reporter gene cassette, depending on their relative locations within the Trappc9 locus (Figures 6A,B). We performed reporter gene assays in cultures of mouse primary hippocampal neurons and embryonic fibroblasts. Compared to the promoter-only construct, regulatory elements Reg-B and Reg-E had significant silencing effects in fibroblasts, but not neurons, indicating a tissue-/cell type-specific function (Figure 6C). Reg-D displayed silencing activity in both neurons and fibroblasts. Unexpectedly, none of the regulatory elements showed enhancer activity in our assay conditions. All three silencing elements are located on the Trappc9-proximal side of the Peg13 DMR and CTCF-binding site (Figure 6A). These silencer elements might contribute to the regulation of tissue-specific expression of Trappc9, Chrac1 and/or Ago2 in vivo. Reg-D might also contribute specifically to the reduced transcription of their paternal alleles in brain, thereby generating an imprinted expression bias in this tissue, although any allele-specific mechanism remains to be elucidated.
Our findings of allelic expression biases of these imprinting cluster genes in newborn mouse brain are in line with previous data from whole transcriptome studies, showing strong (∼90%) paternal and maternal preferences for Peg13 and Kcnk9, respectively, while Trappc9 and Ago2 displayed a more moderate (∼75%) preference for the maternal allele (Babak et al., 2015; Bonthuis et al., 2015; Crowley et al., 2015; Perez et al., 2015; Bouschet et al., 2016; Andergassen et al., 2017; Huang et al., 2017). Chrac1 fell below the threshold of 70% bias in our data and also showed a weaker maternal bias than Trappc9 and Ago2 in the studies by Crowley et al. (2015) and Perez et al. (2015). Brain-specificity of Trappc9 and Ago2 imprinting was also confirmed, since we found expression in kidney to be equal bi-allelic, in line with previous data. Unexpectedly, we did not detect an imprinted expression bias for Trappc9 and Ago2 in cultured hippocampal NSCs (neurospheres), where both genes showed equal bi-allelic expression in bulk sample analysis. By contrast, Peg13 retained its strong paternal expression bias in neurospheres. These findings are reminiscent of another imprinted gene, Dlk1, which loses its mono-allelic paternal expression and becomes bi-allelically expressed in postnatal NSCs of the ventricular zone and hippocampal subgranular zone (Ferron et al., 2011; Montalban-Loro et al., 2021). The change to bi-allelic expression of Dlk1 is associated with gain of methylation at its germline DMR and is a requirement for normal postnatal and adult neurogenesis. However, we did not find any change in methylation at the Peg13 germline DMR in neurospheres. Also, the promoter CGI of Trappc9 remained unmethylated on both alleles, while the second CGI at exon 2 retained its high levels of methylation in neurospheres. Similarly, the Ago2 promoter CGI remained unmethylated on both alleles in neurospheres. Thus, the regulation of allelic expression of Trappc9 and Ago2 in NSCs is likely to differ from that in differentiated neural cells and might involve changes in histone modifications, transcription factor binding and/or enhancer access. In any case, a relevance of Trappc9 expression in NSCs is implied by the finding of reduced numbers of Sox2-positive stem cells in the subventricular zone and hippocampal subgranular zone of knock-out mice (Usman et al., 2022), which might be linked to their microcephaly phenotype (Ke et al., 2020; Liang et al., 2020; Wilton et al., 2020; Aslanger et al., 2022). Due to advances in technology, especially single-cell transcriptomics and highly sensitive in situ hybridization methods, it has now become possible to investigate imprinted gene expression on the cellular level (Varrault et al., 2020; Martini et al., 2022). Instead of a single-cell RNA-seq approach, we used the sc-GEM method (Lorthongpanich et al., 2013; Cheow et al., 2015; Cheow et al., 2016) for a more limited analysis of the genes of this imprinting cluster in single cultured NSCs and differentiated neurons. Our data for Trappc9 and Ago2 show a broad variability of allelic expression status in individual neurosphere cells. All categories of allelic expression, ranging from mono-allelic maternal to equal bi-allelic to mono-allelic paternal and intermediately biased bi-allelic states, were found in significant numbers of cells. Taking into account all single NSCs analyzed, the allelic expression of these two genes leveled out in line with the bulk neurosphere data, i.e., overall there was no allelic bias in the neural stem cell population. Our findings were similar in single neurons that were differentiated from the neurospheres; significant numbers of cells were found for each category of allelic expression. When comparing the cell population of NSCs with the neuronal population for Ago2 expression biases, a slight overall shift from paternal to maternal allelic biases was observed. For Trappc9, the proportion of cells with equal bi-allelic expression was reduced in the neuron population compared to the NSC population, but the proportions of cells with maternal and paternal expression biases, respectively, remained balanced. Neither of the two genes displayed an overall maternal expression preference in the population of single neurons, which is in contrast to the data obtained from brain lysates. However, our neuron culture is not fully representative of all the cell types that would be included in a brain tissue lysate, since we actively selected against dividing glial cells by adding Ara-C to the culture. Furthermore, since the neurons were differentiated from hippocampal NSCs, our culture is likely to contain only a limited range of neuronal cell types. Data from the Allen Brain Map (https://portal.brain-map.org/) in situ hybridization atlas and single-cell transcriptomics indicate medium levels of Trappc9 and Ago2 expression in many neurons of the cortex and hippocampus, with lower levels occurring in some types of neurons as well as astrocytes and oligodendrocytes. Limited histological analysis for Trappc9 by Ke et al. (2020) support these data. On a more general note, we have no indication that our single-cell data are affected by potential technical issues, for example, allele drop-out during reverse transcription, and we have not found a way to test for such eventualities at extremely low numbers of RNA molecules. However, this would affect the weakly expressed allele, i.e., the paternal allele of Trappc9 or the maternal allele for Peg13 as judged from brain lysates, more than the strongly expressed allele and should lead to an increased number of cells with mono-allelic maternal expression of Trappc9, or exclusively cells with mono-allelic paternal expression of Peg13. There is no indication for such an effect in our data. On the contrary, our data show the opposite, i.e., a surprisingly large number of cells that display a biased paternal expression of Peg13 with weak expression of the maternal allele readily detectable, and even cells with predominantly maternal Peg13 expression. Such results would not be expected, if there were a significant rate of allelic drop-out of the weakly expressed allele. Furthermore, our findings of varying mono- or bi-allelic expression states of Ago2 in individual neurons is in line with in situ hybridization data obtained by Bonthuis et al. (2015), who analyzed nascent transcripts in nuclei of brain sections and determined that 46% of Ago2 expressing cells in the arcuate nucleus, and 63% in the dorsal raphe nucleus, showed mono-allelic expression with the remainder having two visible sites of nuclear transcription. For Peg13, our single-cell data indicated a predominantly paternal expression bias in NSCs, and even more so in differentiated neurons, although there was a substantial proportion of cells with paternally biased bi-allelic (instead of mono-allelic) expression. A small number of cells, mainly NSCs, deviated from this expected bias and showed equal bi-allelic or even mono-allelic maternal expression. Although surprising, these findings are not unprecedented. A recent study of imprinted gene expression in single cortical cells identified similar variability and occasional deviations from expected biases (Laukoter et al., 2020). For example, Meg3 (also known as Gtl2), which usually has a strong maternal expression bias, was found to be bi-allelic in a small number of cortical cells. For two other imprinted genes with an expected paternal expression bias, i.e., Inpp5f and Impact, a substantial number of cortical cells deviated towards bi-allelic or predominantly maternal expression (Laukoter et al., 2020). The mechanisms and reasons behind such variable allelic expression states of imprinted genes in individual cells are largely unclear. The cases of Dlk1 gain of methylation and loss of imprinting in NSCs (Ferron et al., 2011; Montalban-Loro et al., 2021), or Grb10 alternative promoter usage on the maternal and paternal alleles (Yamasaki-Ishizaki et al., 2007; Sanz et al., 2008; Garfield et al., 2011) are unlikely models for our findings. On the other hand, random mono-allelic expression (RMAE) effects, especially transcriptional bursting (Reinius and Sandberg, 2015; Chess, 2016; Xu et al., 2017; Varrault et al., 2020), might affect imprinted genes and be the underlying reason for the variable allelic expression states we find in single cells. Such RMAE effects might be stochastic and dynamic, rather than permanent as in the case of random allelic exclusion of immunoglobulin genes, and might involve relatively short-lived CTCF-cohesin chromatin loops (Gabriele et al., 2022). For the human PEG13-KCNK9 locus, CTCF-cohesin binding sites, chromatin looping and enhancer interactions with those two gene promoters have been described (Court et al., 2014). CTCF-binding on the unmethylated paternal allele of the Peg13 gDMR is conserved in mice (Singh et al., 2011; Prickett et al., 2013), but tissue-specific enhancer elements and chromatin looping might differ in this species and could underly the imprinted expression of Trappc9, Chrac1 and Ago2 in murine brain tissue. Our ENCODE3-based search for brain-specific regulatory elements considered active histone and open chromatin marks and resulted in several candidate regions, but unexpectedly these did not show enhancer function when tested in transfected primary neurons or fibroblasts. Instead, two elements had silencing activity specifically in fibroblasts, while a third element silenced reporter gene activity in both neurons and fibroblasts. Typically, active enhancers are associated with active chromatin marks, e.g., H3K27ac, H3K4me1, H3K9ac, and silencers are variably marked by H4K20me, H3K9me3 (typical for heterochromatin and methylated DNA) and/or H3K27me3, although chromatin at silencer regions is still expected to be open for binding of repressive transcription factors and, therefore, also associated with H3K79me2 and H3K36me3 marks (Pang and Snyder, 2020). However, there is currently no widely accepted consensus for a silencer chromatin signature and many silencer elements might be bifunctional elements acting through various mechanisms (Segert et al., 2021). Furthermore, a recent functional study of ENCODE3 candidate cis-regulatory elements (cCREs) found that the majority of annotated cCREs had no effect on transcription, while similar numbers of the remaining elements had enhancer or repressor activity, respectively, which was surprising given that cCREs are predicted to be enhancers (Martinez-Ara et al., 2022). Further experiments will be required to determine whether the silencer elements we identified within the Trappc9 locus might function in an allele-specific way and contribute to the brain-specific imprinted expression bias of this gene. A number of Trappc9 transcript variants have been annotated on ENSEMBL, including two alternative promoters and truncated transcripts, one of which has been described as specifically expressed from the paternal allele in RNA-seq data (Gregg et al., 2010; Hsu et al., 2018). Our attempts to confirm such truncated transcript versions with primer combinations that span transcript-specific and shared exons were unsuccessful. Trappc9 and Peg13 are transcribed in the same direction and Peg13 is an unspliced long non-coding RNA located in intron 17 of Trappc9. Peg13 transcription from the paternal allele might extend further downstream than is currently known, similar to other long non-coding RNAs at imprinted loci, e.g., Nespas, or Meg3/Gtl2 (Ferguson-Smith, 2011; Peters, 2014). Therefore, allelic expression analysis of Trappc9 RNA downstream of the Peg13 start site requires careful consideration. We also found no evidence for a second promoter of Trappc9. The transcriptional start sites we identified at the first non-coding exon will create transcripts with a translational start site in exon 2, which encodes the well-conserved NH2-terminal end of the protein. This would be missing upon alternative promoter usage. From our data, we can exclude a second promoter, which is also of relevance in the context of the variable allelic expression of Trappc9 discussed above. We can exclude a second promoter as a possible explanation for the variable allelic expression in single cells. Overall, the mechanisms of brain-specific imprinted expression of Trappc9 in mice remain to be fully elucidated, but this allelic bias is of biological relevance, since maternal transmission of a knock-out mutation of Trappc9 results in phenotypes similar to homozygous deletion and, vice versa, mice carrying a paternally transmitted mutation are not different from wild-types (Liang et al., 2020). | true | true | true |
|
PMC9596792 | Chao Wang,Yaqiong An,Zhaohua Xia,Xuezhi Zhou,Haibo Li,Shuang Song,Lexi Ding,Xiaobo Xia | The neuroprotective effect of melatonin in glutamate excitotoxicity of R28 cells and mouse retinal ganglion cells | 12-10-2022 | melatonin,glaucoma,nmda,glutamate,oxidative stress,retinal ganglion cell | Glaucoma is the leading cause of irreversible blindness. The progressive degeneration of retinal ganglion cells (RGCs) is the major characteristic of glaucoma. Even though the control of intraocular pressure could delay the loss of RGCs, current clinical treatments cannot protect them directly. The overactivation of N-methyl-D-aspartic acid (NMDA) receptors by excess glutamate (Glu) is among the important mechanisms of RGC death in glaucoma progression. Melatonin (MT) is an indole neuroendocrine hormone mainly secreted by the pineal gland. This study aimed to investigate the therapeutic effect of MT on glutamate excitotoxicity of mouse RGCs and R28 cells. The Glu-induced R28 cell excitotoxicity model and NMDA-induced retinal injury model were established. MT was applied to R28 cells and the vitreous cavity of mice by intravitreal injection. Cell counting kit-8 assay and propidium iodide/Hoechst were performed to evaluate cell viability. Reactive oxygen species and glutathione synthesis assays were used to detect the oxidative stress state of R28 cells. Retina immunofluorescence and hematoxylin and eosin staining were applied to assess RGC counts and retinal structure. Flash visual-evoked potential was performed to evaluate visual function in mice. RNA sequencing of the retina was performed to explore the underlying mechanisms of MT protection. Our results found that MT treatment could successfully protect R28 cells from Glu excitotoxicity and decrease reactive oxygen species. Also, MT rescued RGCs from NMDA-induced injury and protected visual function in mice. This study enriches the indications of MT in the treatment of glaucoma, providing practical research ideas for its comprehensive prevention and treatment. | The neuroprotective effect of melatonin in glutamate excitotoxicity of R28 cells and mouse retinal ganglion cells
Glaucoma is the leading cause of irreversible blindness. The progressive degeneration of retinal ganglion cells (RGCs) is the major characteristic of glaucoma. Even though the control of intraocular pressure could delay the loss of RGCs, current clinical treatments cannot protect them directly. The overactivation of N-methyl-D-aspartic acid (NMDA) receptors by excess glutamate (Glu) is among the important mechanisms of RGC death in glaucoma progression. Melatonin (MT) is an indole neuroendocrine hormone mainly secreted by the pineal gland. This study aimed to investigate the therapeutic effect of MT on glutamate excitotoxicity of mouse RGCs and R28 cells. The Glu-induced R28 cell excitotoxicity model and NMDA-induced retinal injury model were established. MT was applied to R28 cells and the vitreous cavity of mice by intravitreal injection. Cell counting kit-8 assay and propidium iodide/Hoechst were performed to evaluate cell viability. Reactive oxygen species and glutathione synthesis assays were used to detect the oxidative stress state of R28 cells. Retina immunofluorescence and hematoxylin and eosin staining were applied to assess RGC counts and retinal structure. Flash visual-evoked potential was performed to evaluate visual function in mice. RNA sequencing of the retina was performed to explore the underlying mechanisms of MT protection. Our results found that MT treatment could successfully protect R28 cells from Glu excitotoxicity and decrease reactive oxygen species. Also, MT rescued RGCs from NMDA-induced injury and protected visual function in mice. This study enriches the indications of MT in the treatment of glaucoma, providing practical research ideas for its comprehensive prevention and treatment.
Glaucoma is the leading cause of irreversible blindness in the world, and it is characterized by progressive degeneration of retinal ganglion cells (RGCs) and their axons, accompanied by visual field defects (1). The global prevalence of glaucoma in the 40- to 80-year-old population is estimated to be 3.5%. With the number and proportion of the elderly population increasing, 111.8 million people are expected to suffer from glaucoma by 2040 (2). At present, the management of glaucoma mainly focuses on the regulation of intraocular pressure (IOP) and slowing its progress (3). Many studies have shown that even controlling the increase in IOP cannot prevent the death of RGCs and progressive visual field defects (4–6). For patients with end-stage glaucoma, there is no effective neuroprotective method. Therefore, seeking effective optic neuroprotective medication for the treatment of glaucoma is necessary. As an excitatory neurotransmitter, glutamate (Glu) exists widely in retinal neurons and is involved in the signal transmission between photoreceptors, bipolar cells, and RGCs through N-methyl-D-aspartic acid (NMDA) receptors (7, 8). In the pathological state of glaucoma, excess Glu between synapses cannot be effectively removed and can cause NMDA receptor overactivation, calcium overload in nerve cells, and oxidative stress damage, leading to the death of RCGs and degeneration, which is called glutamate excitotoxicity (8). Many studies have confirmed glutamate excitotoxicity to be among the important mechanisms of RGC death in glaucoma progression (9–11). Melatonin (N-acetyl-5-methoxytryptamine, C13N2H16O2) is an indole neuroendocrine hormone mainly secreted by the pineal gland (12). The secretion of MT has a circadian rhythm. After night falls, the synthesize of MT increases, and the secretion level of MT in the body also increases accordingly, reaching a peak at 2-3 am in the morning. The level of MT at night directly affects the quality of sleep. As a classic antioxidant, MT can protect against oxidative stress damage through different mechanisms, including direct scavenging of reactive oxygen species (ROS), regulation of signaling pathways against oxidative stress, and upregulation of glutathione synthesis (GSH). In addition to MT, many of its metabolites also function as ROS scavengers (13). Through different mechanisms, MT can attenuate oxidative stress damage in lipids, proteins, DNA, and many tissues. Besides being secreted from the pineal gland, MT has also been found to be synthesized and released by many ocular structures, including the retina, ciliary body, lens, and Harderian gland in chickens (13, 14). Studies have shown that MT can exert neuroprotective effects through its anti-oxidative stress effect (15–17),, although its role in neuroprotective effects in glaucoma is still unclear. In this study, we found that MT showed an effective neuroprotective effect against neuronal glutamate toxicity, and it significantly reduced NMDA-induced loss of RGCs. Moreover, MT treatment significantly reversed changes in the retinal transcriptome caused by NMDA. All these results highlight the potential value of MT as a potential medication for neuroprotection treatment in glaucoma.
Immortalized R28 cells (Key Laboratory of Ophthalmology, Xiangya Hospital, Central South University, Changsha, China) were maintained in low-glucose Dulbecco’s modified Eagle’s medium (11885084, Gibco, Carlsbad, USA) supplemented with 10% fetal bovine serum (FSP500, ExCell Bio, Jiangsu, China) and 1% penicillin-streptomycin (C100C5, NCM Biotech; Zhejiang, China) at 37°C with 5% CO2. In the glutamate excitotoxicity model, cells were treated with L-Glutamate (ab120049, Abcam, Cambridge, UK) and incubated for 24 h.
Cell viability was measured by cell counting kit-8 (CCK-8; C6005, NCM Biotech). R28 cells were seeded in 96-well plates at a density of 5000 cells/well and cultured in a medium containing various concentrations of Glu (5, 10, 15, 20, 25 mM). After 24 h incubation, 10% CCK-8 was added and incubated at 37°C for 3h, as per the manufacturer’s instructions. Absorbance was measured at 450 nm using a microplate reader. Meanwhile, propidium iodide (PI)/Hoechst was applied to calculate the R28 cell survival rate after Glu and MT (M5250, Sigma-Aldrich, St. Louis, MO, USA) were treated for 24 h; cells were stained using apoptosis and necrosis assay kit (C1056, Beyotime, Shanghai, China) and pictured using an optical microscope (Eclipse C1, Nikon, Tokyo, Japan).
Intracellular ROS were detected using a cellular ROS assay kit (ab113851; Abcam). The collected cells were digested with trypsin and then stained in culture media with 20 µM DCFDA and incubated for 30 minutes at 37°C. The cells were washed with 1× buffer after incubation and analyzed immediately with a flow cytometer. Forward and side scatter gates were established to exclude debris and cellular aggregates from the analysis. DCF was excited by the 488 nm laser and detected at 535 nm (typically FL1). The mean florescence intensity (MFI) were analyzed by Flowjo software version 10.0.7.
A micro reduced GSH assay kit (BC1175, Solarbio, Beijing, China) was used to detect reduced GSH. A total of 5 million cells were collected and cleaned twice with PBS. The GSH extract was then added twice for repeated freeze–thaw (frozen in liquid nitrogen and dissolved in a 37°C water bath) and centrifuged at 8000 g for 10 min, and the supernatant was collected at 4°C. The GSH content was detected according to the instructions and standardized according to the number of cells.
C57BL/6 mice (8 weeks old; Slaccas, Changsha, China) were fed with standard laboratory food and water in a comfortable environment with a 12 h light–dark cycle. All the experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Central South University (Changsha, China). All mice were divided into Three groups: Sham (only acupuncture without injection), NMDA (20 mM), and MT (20 mM NMDA + 400 mM MT). All the mice were anesthetized with pentobarbital (1%, 80 mg/kg, intraperitoneal injection; Beijing Sanshu, China) and then operated on under a stereomicroscope. Oxybuprocaine hydrochloride (Santen Pharmaceuticals, Tokyo, Japan) was used to induce ocular surface anesthesia, and tropicamide phenylephrine (Santen Pharmaceuticals) was used to dilate the pupils. A 30 G needle was inserted into the vitreous cavity along the limbus and injected at a volume of 1 µL per eye. Tobramycin dexamethasone eye ointment (Alcon Inc, Geneva, Switzerland) was used to prevent infection after injection. The mice were euthanized 5 d after the injection, and their eyeballs were removed with tweezers for the follow-up research.
Visual function was assessed by flash visual-evoked potential analysis (FVEP) 5 d after intravitreal injection, all after anesthesia. After 15 min of dark adaptation, the following 3 electrodes were fixed separately and inserted under the skin: ground electrode (ack), cathode (anterior bregma), and anode (occipital bone). After covering the contralateral eye, the images of both eyes were measured by a multifocal electroretinography recorder (GT-2008V-VI, Gotec, Chongqing, China) and recorded by Ganzfeld electrodiagnostic system (Gotec). The time of the flash is 100 ms. The first negative wave amplitude and first positive wave latencies were used to assess the visual function in mice.
The mice were euthanized 5 days after modeling, and their eyeballs were removed and fixed with an FAS eyeball fixator (G1109, Servicebio, Wuhan, China). The eyeballs were embedded in paraffin and cut into 4 μm vertical sections. Sections were stained with hematoxylin and eosin (H&E; G1120, Servicebio) according to the manufacturer’s instructions and visualized using an optical microscope (Eclipse C1). CaseViewer software (3DHISTEC, Sysmex, Switzerland) was used to measure the thickness of the ganglion cell layer at distances of 300, 600, 900, 1200, and 1500 μm from the optic nerve center.
The mice were euthanized 5 days after modeling, and their eyeballs were removed and fixed with 4% paraformaldehyde (G1101, Servicebio) fixation for 1 h, and the retinas were detached under a stereomicroscope. The retinas were sealed with 5% bovine serum albumin and 0.5% Triton-X100 in PBS for 1.5 h at room temperature, followed by incubation with primary antibody RBPMS (ab152101, Abcam) at 4°C overnight. They was cleaned with 0.5% Triton-X100 in PBS 4 times for 5 min each were then incubated with fluorescent-labeled secondary antibody away from light for 2 h at ambient temperature. The retinas were then viewed and pictured by optical microscope (Eclipse C1).
The mouse retinas were collected 5 days after NMDA intervention. Three individual retinas were treated as one sample, and each group contained 3 samples. RNA was isolated by total RNA kit (R6834-01, Omega Bio-Tek),. Total amounts and integrity of RNA were assessed using the RNA Nano 6000 assay kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). After RNA was converted to cDNA, the samples were sequenced by the Illumina NovaSeq 6000 at Novogene (Beijing, China). Genes with a fold-change ≥1.5 identified by edgeR and a false discovery rate <0.05 were considered differentially expressed (BMKCloud, http://www.biocloud.net/). Gene functional annotations were based on the Kyoto Encyclopedia of Genes and Genomes (KEGG, https://www.genome.jp/kegg/) and Gene Ontology (GO, http://www.geneontology.org/) databases.
SPSS 26.0 statistical software (IBM Corp., Armonk, NY, USA) was used for statistical analysis of all data. All data were presented as the mean ± standard deviation (SD). One-way analysis of variance (ANOVA) was used to assess the significance differences of cell viability, ROS, GSH, RGCs survival and FVEP results between groups. Repeated measures ANOVA was used to assess thickness of retinal ganglion cell complex (GCC). Charts were built using GraphPad Prism 6.0 (GraphPad Inc., La Jolla, CA, USA). P value <0.05 was statistically significant.
To investigate the appropriate concentration of Glu, R28 cells were treated with 5-25mM Glu at different concentrations for 24 h. CCK-8 assay results showed that cell viability decreased gradually with increasing Glu concentration in a concentration-dependent manner. Compared with the control group, 10 mM (47.90 ± 15.50%), 15 mM (26.41 ± 5.48%), 20 mM (5.41 ± 3.86%), 25 mM (4.73 ± 1.43%) significant decreased cell viability with Glu treatment for 24 h. (P <0.001, n = 4) ( Figure 1A ). In subsequent experiments, R28 cells were treated with 10 mM Glu for 24 h as the immobilization condition. Subsequently, we investigated the protective effect of different concentrations of MT on glutamate-induced excitotoxicity injury. The results showed that compared with the Glu group, the cell viability of the MT group was significantly increased with the increasing MT concentration. The cell viability reached 109.1 ± 6.9% when the concentration of MT was at 400 μM (P <0.001, n = 6) ( Figure 1B ), suggesting that MT has a protective effect on Glu-induced R28 cell damage, and the MT at concentration of 600 μM, 800 μM and 1000 μM also showed good protective effect (P <0.001, n = 6). Meanwhile, PI/Hoechst staining was used to confirm this view, and it was observed that R28 cells died more after 24 h of glutamate treatment, while MT saved this damage ( Figure 1C ). These results suggest that MT can protect R28 cells from glutamate-induced excitotoxicity.
To investigate the effects of MT on Glu-induced oxidative stress in R28 cells, intracellular ROS and reduced GSH levels were detected. The results showed that ROS levels increased gradually over time after Glu treatment and peaked at 12 h (7.42 ± 0.52), and MT treatment could ameliorate the changes induced by Glu (3.67 ± 0.30, P <0.001, n = 3) ( Figures 2A, B ). At 24 h, MT was still protective (5.14 ± 0.43, P <0.01, n = 3), but this was not as significant as at 12 h ( Figures 2A, B ). Meanwhile, with the increase in Glu treatment time, the intracellular GSH level gradually decreased and reached its lowest at 24 h (0.20 ± 0.01, P <0.001, n = 3) ( Figure 2C) . However, the GSH level of the MT group was not significantly improved when compared with the Glu group (0.16 ± 0.02, P >0.05, n = 3) ( Figure 2C ).
To further determine the protective effect of MT on retinal excitotoxicity, the thickness of GCC was measured after H&E staining, and RGCs were counted and quantitatively analyzed after being labeled with RBPMS by retinal immunofluorescence. H&E staining showed that the retinal GCC thickness of mice in the NMDA group was significantly thinner than that in the control group 5 days after intravitreal injection of NMDA (P <0.001, n = 4) ( Figures 3A, B ). MT treatment could effectively inhibit the thinning of the GCC layer caused by NMDA at 300, 600 and 900 μm from the optic nerve center (P <0.05, n = 4). Retinal immunofluorescence showed that the density of RGCs in the NMDA group was significantly lower than that in the control group, while the number of surviving RGCs in the MT group (1883.10 ± 124.63) was significantly better than that in the NMDA group (849.30 ± 47.10) but still lower than that in the control group (2694.60 ± 145.85, P <0.001, n = 4) ( Figure 3C ). These results suggest that MT has a protective effect on NMDA-induced retinal injury in mice.
We also studied the effect of MT on electrophysiological activity of the retina and its protective effect on visual function in mice. The amplitude of N1 wave was decreased NMDA treatment (1.98 ± 0.76 µV) compared with control group (4.87 ± 1.10 µV), and MT increase amplitude of N1(4.85 ± 0.82 µV, P <0.001, n = 6).The latencies of P2 wave were prolonged 5 days after intravitreal injection in the NMDA group (115.42 ± 5.45 ms) compared with control group (84.83 ± 3.52 ms), and MT ameliorated this change (103.91 ± 5.28, P <0.005, n = 6) ( Figure 4 ). These results suggest that NMDA causes retinal dysfunction in mice, and MT can improve the visual conduction dysfunction induced by excitotoxicity.
To investigate further the mechanism of the neuroprotective effects of MT on the retina, RNA sequencing analysis was performed. Compared with the control group, the NMDA-treated group had 1519 upregulated genes and 1663 downregulated genes. With the intervention of the MT, 139 genes were upregulated and 227 genes downregulated ( Figure 5A ). MT treatment mitigated the expression of approximately 49 upregulated genes and 57 downregulated genes induced by NMDA ( Figure 5B ). These genes included 5 neuroactive ligand-receptor interaction-related genes (Oprl1/Ptafr/Adcyap1r1/Lpar6/Crhr1), 3 PI3K-Akt signaling pathway-related genes (Col6a3/Lpar6/Gng4), and 3 calcium signaling pathway-related genes (Ptafr/Prkcg/Orai3). These results suggest that MT exhibits its neuroprotective effect by ameliorating retinal transcriptome abnormalities ( Figure 5C ).
KEGG and GO analyses were performed to densify the signaling pathway and biological process changes in the retina. We performed KEGG analysis on the differentially expressed genes in the control, NMDA, and MT groups. This indicated that the PI3K-Akt and MAPK signaling pathways were both crucial after NMDA intervention. After MT treatment, the PI3K-Akt and JAK-STAT pathways were involved in rescuing the injury induced by NMDA ( Figure 6A ). GO analysis results showed that the retinal biological process, cellular component, and molecular function were all altered by MT intervention. The differentially expressed genes were enriched in the biological process ( Figure 6B ).
The characteristic death of RGCs is one of the most important features of glaucoma, which can cause irreversible visual field defects and seriously affect the life quality of patients (18, 19). Although many glaucoma medications have been applied in clinical treatment, their use for glaucomatous neuroprotection is still very limited, and there are no clear clinical outcomes (20, 21). Many studies have shown a potential relationship between MT and glaucoma. Patients with glaucoma are often accompanied by sleep disturbances, anxiety, and depression, and studies have shown that glaucoma is also associated with disturbances in the rhythm of MT secretion (22, 23). Recent studies showed that urinary 6-sulfatoxymelatonin, the main metabolite of serum MT in glaucoma patients, is significantly lower than normal, suggesting the possibility of a circadian rhythm disturbance in glaucoma patients, which MT can restore (24). Focusing on the eye, although MT can be secreted by various eye structures, and the aqueous humor also contains a certain concentration of MT (25), the role of MT in the eye is still unclear. Concentrations of MT have been shown to be 3 times higher in aqueous humor in patients with elevated IOP than in normal patients, and the same was also observed in a mouse model of glaucoma (26). Preoperative treatment with oral MT has been shown to reduce IOP in patients who have undergone cataract surgery (27). Animal models and clinical trials have also shown that MT and its analogs can reduce IOP (28, 29). Some studies have demonstrated that MT exerts antiapoptotic and neuroprotective effects on retinal neurons after hypoxia-ischemia and acute intraocular hypertension (30, 31). In our study, we established the classical NMDA-induced retinal injury model, which imitated the different mechanisms of RGC death in the pathogenesis of glaucoma. We found that MT has a significant protective effect on cellular Glu excitotoxicity both in vivo and in vitro and provides a supplement to the protective role of melatonin in the pathogenesis of different glaucoma. We found that MT at a concentration of 400 μM had a 100% protective effect on Glu-induced cell excitotoxicity in R28 cells, and a high concentration of 1000 μM had no toxic effect on cells, confirming that MT is effective and safe. MT also showed a good neuroprotective effect in vivo. MT rescued NMDA-induced RGC loss and GCC thinning, and through the detection of FVEP, it was confirmed that melatonin can restore partial visual function in mice. As an endocrine hormone with strong anti-oxidative stress ability, MT has strong potential for the neuroprotection of glaucoma. As a classic antioxidant, MT can effectively scavenge ROS and increase the content of intracellular GSH to resist oxidative stress injury (32). In our study, we found that after glutamate excitotoxicity injury, although MT had a significant protective effect on R28 and RGCs and a significant recovery of visual function in mice, it did not increase the content of GSH. As a common antioxidative product, its depletion is much greater than its synthesis in the glutamate excitotoxicity process. ROS was significantly higher than that in the Glu group at 24 h, but it still had a 5.16-fold increase compared to the control group. Therefore, we speculate that, in addition to the scavenging effect of MT on ROS, other mechanisms also play a key role in the process of neuroprotective effect. To explore the mechanism of action of the neuroprotective effect of MT, we conducted RNA sequencing of the retina, and the sequencing results of this study found that MT rescued abnormal retinal transcriptome expression induced by NMDA. Through further gene enrichment, we found significant changes in the PI3K-AKT and MAPK signaling pathways in NMDA-induced retinal injury, and after MT treatment, different genes were enriched to the P13K-AKT and JAK-STAT signaling pathways. This is consistent with many of the studies that showed PI3K-AKT and MAPK to be involved in the occurrence and development of glaucoma and play an important role in the death of RGCs (33–36). Studies have also shown that RGCs are protected by intervening PI3K-AKT and JAK-STAT signaling pathways through regulating apoptosis, autophagy, and oxidative stress processes (37–41). In our study, sequencing results showed a neuroprotective effect, suggesting that MT may depend on the above pathways. However, the specific mechanism of action needs to be studied further.
This study explored the neuroprotective effects of MT on NMDA-reduced RGC death and Glu-induced R28 cell excitotoxicity. It found that MT successfully rescues RGCs from NMDA-reduced injury and protects visual function in mice. MT also protects R28 cells from Glu excitotoxicity and decreases ROS. RNA sequencing indicated that MT treatment repairs the abnormal transcriptome caused by NMDA, and PI3K-AKT and JAK-STAT signaling pathways may play an important role in this process. This study provides practical research ideas for the comprehensive prevention and treatment of glaucoma.
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found here: https://www.ncbi.nlm.nih.gov/bioproject/, SRA PRJNA867307.
The animal study was reviewed and approved by Institutional Animal Care and Use Committee (IACUC) of Central South University.
CW and YA wrote the first draft of the paper. XX and LD edited the paper. XX, LD and CW designed research. CW, YA performed laboratory research. ZX performed FVEP examination. XZ performed bioinformatics Analysis, CW, YA and SS analyzed data. All authors contributed to the article and approved the submitted version.
This study was financially supported by The National Key Research and Development Program of China (No.2020YFC2008205), The National Natural Science Foundation of China (No. 82171058, No.81974134 for XX, No.82070966 to LD, No.8210041510 to CW), Key R&D Plan of Hunan Province of China (No.2020SK2076 to XX), Science and Technology Innovation Program of Hunan Province (No.2021RC3026 to LD), Natural Science Foundation of Hunan Province (No.2021JJ41021 to CW) and China Postdoctoral Science Foundation (No.2021M693556 to CW).
We thank Lemeng Feng, Weizhou Fang, Shirui Dai, Cheng Zhang and Wulong Zhang for technical assistance. We thank Scribendi (https://www.scribendi.com/) for editing the English text of a draft of this manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. | true | true | true |
|
PMC9596927 | Mahsa Yousefian,Abdolhamid Angaji,Elham Siasi,Seyed Ali Rahmani,Shamsi Abbasalizadeh Khiaban | Role of CYP1A1, CYP2D6, and NOS3 gene polymorphisms in idiopathic recurrent pregnancy loss in the Iranian Azeri population: A case-control study | 06-09-2022 | Recurrent pregnancy loss,Polymorphism,CYP1A1,CYP2D6,NOS3. | Abstract Background It is estimated that 1-5% of couples suffer from recurrent pregnancy loss (RPL). Recent studies have shown the effects of gene polymorphisms in RPL. Objective The aim of this study was to evaluate 3 gene polymorphisms including rs1048943 of CYP1A1, rs28371725 of CYP2D6, and rs7830 of NOS3 in idiopathic RPL to identify their association with RPL. Materials and Methods Blood samples were collected from 136 women with at least 2 consecutive idiopathic miscarriages (case group) and 136 women with no history of miscarriage and at least one successful pregnancy (control group) from the Iranian Azeri population. This study was carried out between April 2018-April 2020. Amplification-refractory mutation system polymerase chain reaction was used for the rs7830, rs1048943 and rs28371725 polymorphisms in order to genotype each extracted genomic DNA sample. After that, Chi-square, Fisher's exact test and logistic regression were used to investigate whether each of these polymorphisms is associated with RPL. Results Among these polymorphisms, only rs1048943 of CYP1A1 showed a statistically significant association with RPL in the Iranian Azeri women studied. Conclusion Our results suggest that CYP1A1 gene polymorphisms might be associated with a reduced risk of RPL. Further studies in other populations and in the same population with a larger sample size, as well as functional genomics analyses such as gene expression analyses or epigenetic studies are required to validate our results. | Role of CYP1A1, CYP2D6, and NOS3 gene polymorphisms in idiopathic recurrent pregnancy loss in the Iranian Azeri population: A case-control study
It is estimated that 1-5% of couples suffer from recurrent pregnancy loss (RPL). Recent studies have shown the effects of gene polymorphisms in RPL.
The aim of this study was to evaluate 3 gene polymorphisms including rs1048943 of CYP1A1, rs28371725 of CYP2D6, and rs7830 of NOS3 in idiopathic RPL to identify their association with RPL.
Blood samples were collected from 136 women with at least 2 consecutive idiopathic miscarriages (case group) and 136 women with no history of miscarriage and at least one successful pregnancy (control group) from the Iranian Azeri population. This study was carried out between April 2018-April 2020. Amplification-refractory mutation system polymerase chain reaction was used for the rs7830, rs1048943 and rs28371725 polymorphisms in order to genotype each extracted genomic DNA sample. After that, Chi-square, Fisher's exact test and logistic regression were used to investigate whether each of these polymorphisms is associated with RPL.
Among these polymorphisms, only rs1048943 of CYP1A1 showed a statistically significant association with RPL in the Iranian Azeri women studied.
Our results suggest that CYP1A1 gene polymorphisms might be associated with a reduced risk of RPL. Further studies in other populations and in the same population with a larger sample size, as well as functional genomics analyses such as gene expression analyses or epigenetic studies are required to validate our results.
Pregnancy loss is defined as a loss of pregnancy before the end of the 20 wk of gestation (1). Recurrent pregnancy loss (RPL) is defined as at least 2 consecutive miscarriages (2-4). It is estimated that 1-5% of couples suffer from RPL (5, 6). Although in 50% of cases several factors such as endocrine dysfunction, infections, environmental factors, and parental chromosomal abnormalities may cause RPL, in 50% of cases the cause is unknown (7-9). Several studies have examined various polymorphisms of candidate genes that encode different mediators which may affect susceptibility to idiopathic RPL (4, 10, 11). A group of these genes belongs to metabolic enzymes. Genetic polymorphisms of these genes may affect the balance of phase I / phase II detoxification enzymes (12). One of these enzymes is encoded by CYP1A1 which is located on 15q24.1 and includes 7 exons. This gene acts in a 2-step process of detoxifying toxins. In the first step, CYP1A1 is required for activation of toxic components. These components are required for the 2 detoxification step. The polymorphisms of this gene can be directly linked with functional disturbance diseases and conditions like cancers and idiopathic male infertility. A recent study showed that this gene can influence normal estrogen metabolism and placental function (13). It seems that its polymorphisms may also lead to RPL. Another gene of this pathway is CYP2D6 which is located on 22q13.2 and includes 9 exons (14, 15). This gene is important for pharmacogenetics; it is involved in the metabolism of over 150 drugs and has been studied in individuals with suicidal thoughts or depression (16). CYP2D6 has an important role in catalyzing the oxidation of testosterone to androstenediones (17), both of which increase during pregnancy. So, it seems that any change in this enzyme may lead to an increased risk of RPL. There is no autonomic innervation in fetoplacental blood vessels and the regulation of vascular functions at the fetomaternal interface is mediated by endothelial nitric oxide synthase (NOS3), which is a vasoactive mediator (18). This gene is located on 17q36.1 and includes 26 exons. It encodes an enzyme that is important in producing vascular NO (19, 20). “NO is a gaseous molecule, which serves different physiological regulatory functions in the regulation of reproduction, such as the formation of new blood vessels, enhancement of blood supply through the maternal arteries to the placenta, regulation of the placental vessel tones, and immune protection of the fetus. All these factors are required for a successful pregnancy outcome and any disturbance in these steps may increase the risk of miscarriage" (21). In the first trimester of pregnancy, trophoblast cells express a large amount of NOS3, so any polymorphism in coding or noncoding regions of the gene may change its activity or expression level, which may lead to RPL (19). The aim of this study was to determine the relationship between 3 polymorphisms of these genes and RPL in the Iranian Azeri population.
This case-control study was carried out in the Rahmani genetic lab, Tabriz, Iran, during April 2018-April 2020 with 2 groups. The control group (n = 136) consisted of 19-45 yr-old women with no history of miscarriage or infertility and with at least 1 successful pregnancy and a delivery without any complications. The case group (n = 136) consisted of 16-42 yr-old women with at least 2 consecutive idiopathic miscarriages and no successful pregnancies. It should also be considered that all the women in both groups were Iranians with Azeri origin. All women with RPL due to infections, uterine conformational abnormalities, immune disorders, hormonal abnormalities (including thyroid and prolactin disorders) and chromosomal abnormalities in themselves or their spouses were excluded from the study. The sample size was calculated with this formula: N = (Z1-α /2) 2 p (1-p) /d2, (Z1-α /2 = 1.96, p = 0.5, d = 0.1), N = 96. So at least 96 samples were needed to perform our study but in order to control the random sampling of the errors, we increased the sample size to 136 in each group. Please refer to table I for the demographic and clinical characteristics of the cases with RPL and the control group.
5 peripheral blood samples were taken from each individual. Genomic DNA was isolated from a 1 ml ethylenediamine tetraacetic acid-anticoagulated peripheral blood sample using a DNA extraction kit (KBC blood DNA extraction kit, Cat. No. K1135, Tehran, Iran) according to the manufacturers' instructions. The quality of each sample was then estimated using a nanodrop for assessment of nucleic acid purity.
To confirm each polymorphism in its genetic region, the minor allele frequency (MAF) should be more than 0.01. Each single nucleotide polymorphism (SNP) was chosen according to its MAF in the 1000 Genome Project. MAF has been reported as C = 0.133387/668 for rs1048943, T = 0.063498/318 for rs28371725 and T = 0.361821/1812 for rs7830 (22). Rs1048943 is a missense coding sequence variant that causes Ile to Val change in the cyp1a1 protein; rs28371725 is a GA intron variant which causes alternative splicing that can alter cyp2d6 function of the protein; and rs7830 is a GT intron variant that leads to a different transcript. Detailed information about the polymorphisms including the full and abbreviated gene names, residue change and mutation type is presented in table ІІ. Genetic polymorphisms of each participant for rs1048943 and rs28371725 were detected by allele-specific polymerase chain reaction (ARMS-PCR) and the polymorphisms for rs7830 were determined by tetra primer amplification refractory mutation system PCR (T-ARMS-PCR). The ARMS primers were designed using a WASP webpage with the address https://bioinfo.biotec.or.th/WASP/ and for T-ARMS-PCR the designs were done by the primer1 webpage with the address https://http://primer1.soton.ac.uk/. Detailed information about the primers is listed in table ІІІ and table ІV. After their design, all the primers were evaluated with the BLAST-NCBI database and analyzed using the Oligo Analyzer software to ensure the validation of each primer. To increase the specificity of each forward primer in ARMS-PCR and each of the 2 inner primers in T-ARMS-PCR, an extra mismatch was designed in the 2 nucleotide for ARMS-PCR and in the third nucleotide for T-ARMS-PCR from the 3end. Amplifications were carried out in a thermal cycler (Peqlab peqSTAR 96, Erlangen, Germany) with 2 tubes per person using ARMS-PCR, (1 for the wild allele and the other for the mutant allele) with a pair of primers in each tube (wild/mutant forward and common reverse) and 1 tube for each sample using T-ARMS-PCR with 4 primers in each tube. Each tube contained 50 ml: 25 ml of 2PCRBIO Taq Mix Red Master Mix (Cat# PB10.13-02, London, England), 2.0 ml of each primer, 100-500 ng of genomic DNA, and up to 50 ml final volume of deionized water. After pre-denaturation at 95 C for 5 min, the PCR was carried out for 35 cycles of 30 sec at 94 C; 30 sec at 57 C for CYP1A1, 52 C for CYP2D6 and 69.3 C for NOS3; 30 sec at 72 C and at the end of the 35 cycles, the final extension at 72 C for 2 min to complete the extension of all DNA fragments. The PCR products were analyzed in 2% agarose gel stained with gel stain and with a 50 base pairs (bp) DNA ladder as the template of measurement. To ensure the accuracy of the SNP genotypes, 15% of the samples were selected randomly and re-genotyped to verify the initial results. The results confirmed that the genotyping was valid and consistent. The sizes of the PCR products were 220 bp and 133 bp for the rs1048943 and rs28371725 polymorphisms, respectively, and for the rs7830 polymorphism the sizes were 441, 227 and 268 bp for the common, wild and mutant products, respectively. The PCR products are shown in figures 1, 2 and 3.
The Islamic Azad University, Tabriz Branch Ethics Committee, Tabriz, Iran approved this study (Code: IR.IAU.TABRIZ.REC.1398.022). All the participants completed a written informed consent form. The methods were performed in accordance with the ethical principles, national norms, standards, relevant guidelines, and regulations for conducting medical research in Iran.
Central tendency and dispersion of the demographic data related to the clinical characteristics of our case and control groups were examined by descriptive statistics such as mean and standard deviation. Fisher's exact test was used to compare the case and control groups. A p 0.05 was considered statistically significant. For the SNPs with a p 0.05, the Hardy-Weinberg equilibrium analysis was performed by using the Chi-square test and Fisher's exact test (for multiplicative and additive models of rs1048943) to compare the observed and expected genotype frequencies in both the case and control groups to determine the model of association. It was found that the polymorphisms of all of the SNPs followed the multiplicative model. The additive model which is independent of the Hardy-Weinberg equilibrium was evaluated too. To show the effect of these polymorphisms on RPL, odds ratios (OR) with a 95% confidence interval (95% CI) were calculated using logistic regression, Chi-square test and Fisher's exact test (for multiplicative and additive models of rs1048943) in the Statistical Package for the Social Sciences (SPSS), version 26 (SPSS Inc., Chicago, Illinois, USA).
The mean age of the women in the control group was 31.9 6.3 yr (range 19-45) and 28.9 6.3 (range 16-42) in the case group. Three candidate polymorphisms were genotyped and analyzed during our study. A comparison of the frequencies of the polymorphisms in the case and control groups are shown in table V. According to the results, rs28372725 of CYP2D6 (p= 0.75) and rs7830 of NOS3 (p = 0.46)did not show any statistically significant difference between the case and control groups and for this reason no more statistical steps were followed for these 2 SNPs. Regression analyses for rs1048943 of CYP1A1 (p 0.001)under the multiplicative model and additive model were carried out and the results are shown in table VI. According to our findings, rs1048943 of CYP1A1 was the only polymorphism that was statistically associated with RPL in both multiplicative and additive models. As shown in table VI, the association of rs1048943 with RPL (p 0.001, OR [95% CI] = 0.342 [0.221-0.531]) was observed under the multiplicative genetic model. In this SNP, TT is the wild genotype, so it was considered as the reference genotype. According to the additive genetic model, CC (OR [95% CI] = 0.111 [0.032-0.388], p 0.001) and CT (OR [95% CI] = 0.463 [0.265-0.807], p = 0.01) of this SNP were associated with RPL using TT as the reference genotype. Although the genotype frequencies of TT and CT in rs28372725 of CYP2D6 were higher in the controls than in the cases, we failed to find any statistically significant association of these with RPL. Finally, rs7830 of NOS3 did not show any significant association with RPL either.
CYP1A1 plays an important role in the oxidation of polycyclic aromatic hydrocarbons like benzopyrene and polychlorinated biphenyls. These substances are unusual environmental toxicants (13). The role of this enzyme in activating carcinogens in cancers has been shown in previous studies (23-25). A recent study also showed that placental CYP1A1 mRNA levels were higher in women with RPL (26). We also know that CYP1A1 can affect the metabolism of estrogen and normal functions of the placenta (13). In fact, this enzyme participates in the metabolism of estrogen by catalyzing the 2-hydroxylation of stradiol (27) and can convert endogenous estrogens into more hydrophilic compounds (28). This enzyme also has an important role in the metabolism of enzymatic xenobiotics that are specifically induced in women exposed to tobacco smoke. These xenobiotics may transfer through the placenta to the fetus and cause toxicity in the fetus or may affect the expression or production of placental hormones and change the function of proteins (29). Thus, it seems that it could be a good candidate gene in relation to RPL. One of its polymorphisms is rs1048943 AG which is located in 2455 nucleotides and causes IleVal in the amino acid chain. This amino acid change occurs near the hem group of the protein and causes the enzyme activity to double (30). This polymorphism is a risk factor for pharyngeal, prostate, lung, oral, ovarian, bladder, colorectal, and cervical cancers (31). A protective association also was shown between this polymorphism and coronary artery disease in a study conducted with the Han population of China in 2017 (32). In our study, we found that rs1048943 was associated with a reduced risk of RPL in women of Iranian Azeri origin. However, another study done in 2014 in Russia did not show any association between this polymorphism and RPL (33). Given the varied results, studying this common polymorphism of CYP1A1 in larger populations with different genetic backgrounds is recommended. CYP2D6 is active in catalyzing the oxidation of testosterone to androstenediones (17). Both of these hormones increase during pregnancy. The normal fertility process requires an adequate supply of sex hormones and the normal functioning of CYP2D6 is needed to reach this threshold. CYP2D6 is also an important modulator of the detoxification system, and can protect the environment of the utero-placental tissue from being affected by overwhelming oxidative stress (6). The activity of this enzyme begins to increase in the early stages of the 2 trimester of pregnancy and continues to increase as the pregnancy progresses. Nowadays it is also known that many commonly used drugs in pregnancy are metabolized by this enzyme. Considering the highly polymorphic nature of this gene (28), it seems that this could be a good candidate gene in RPL associated studies. This enzyme is also important in pharmacogenetics and is active in the biotransformation of more than 150 drugs including antipsychotics, antidepressants, analgesics, beta-blockers, and antiarrhythmics (16). Rs28371725 causes an A G change in nucleotide 2989 in intron 6 which brings about alternative splicing and so changes the structure of its protein. This structure change results in a reduction in enzyme activity (reduction of metabolizing of beta-blockers). The influence of this polymorphism has been studied in cases of early onset of severe pre-eclampsia; its therapeutic responses and the association between this polymorphism and early-onset preeclampsia were observed (34). To the best of our knowledge, ours is the first study to evaluate the effect of this SNP in RPL. Our research showed that the association between this SNP and RPL in the Iranian Azeri population was not significant. Another study which was conducted in the South of India in 2004 investigated the association between rs3892097 of this gene and RPL but it also did not find any positive or negative association between rs3892097 and RPL (35). “NOS3 encodes an enzyme that generates NO in endothelial cells and is involved in the regulation of vascular functions". The endometrial expression of NOS3 reaches its peak in humans during implantation (36). NO also plays an important role in blood pressure control, cardiovascular homeostasis, the metabolism of glucose, and insulin resistance during pregnancy. A reduction of NO production may cause pregnancy-related vascular problems like RPL. Additionally, alteration in NO synthesis may cause premature labor by affecting the inflammatory response and uterine contractility regulation (37). As a vasoactive agent, NO causes vasodilation and increases the rate of nutrient supply and oxygen perfusion in the umbilical cord. So, any decrease in its rate of production may cause intraurine hypoxia, restriction in fetal growth and increased feto-placental vascular resistance (38). We also know that the inhibition of NO synthesis in pregnant rats causes hypertension, proteinuria, thrombocytopenia and fetal growth restriction that can all lead to miscarriage (39). Hence, this gene could also be a good candidate gene for RPL associated studies. Rs7830 (GT) is associated with 2 different genes: NOS3 and ATG9B, but because it is located within a silencer motif, TGGGGAC, where a GT change can lead to a different transcript, it seems that it influences the NOS3 gene more. Although we did not find any association between this SNP and RPL, the association of this SNP with end-stage renal disease has been previously demonstrated (40). The association between NOS polymorphisms and renal dysfunction (41), atherosclerotic vascular diseases (42), and advanced diabetic nephropathy (43) have also been shown. A study done in 2013 in China could not find any association between this SNP and RPL (44). Another study conducted in Russia in 2019 showed a positive association between rs2070744 of the NOS3 gene with miscarriage (20). This research had some limitations. Firstly, we did not have any information about the women's family history of miscarriage. Secondly, although we asked about the practice of smoking, we did not have any information about other lifestyle habits like alcohol consumption that may have affected the results. Thirdly, our study was limited to only 1 origin within 1 country as we did not assess RPL cases and healthy populations from other ethnic groups or countries. Fourthly, our case and control groups were relatively small, and finally, we investigated only 1 SNP from each gene. Therefore, in future studies, it would be beneficial to use larger groups from various nations, with awareness of the family history of miscarriage and lifestyle practices, and also to investigate other SNPs from these genes.
Our results indicated that while rs1048943 of CYP1A1 wasassociated with a decreased risk of RPL in the studied population, there was no statistically significant association between rs28372725 of CYP2D6 or rs7830 of NOS3 and RPL in the same population. Further studies in various populations are needed to confirm whether these polymorphisms have any positive or negative associations with RPL.
The authors declare that there is no conflict of interest. | true | true | true |
|
PMC9597367 | Marwa A. Ali,Sherin Khamis Hussein,Abeer A. Khalifa,Amani M. El Amin Ali,Marwa S. Farhan,Amal A. Ibrahim Amin,Esam Ali Mohamed | The Ifng antisense RNA 1 (IFNG-AS1) and growth arrest-specific transcript 5 (GAS5) are novel diagnostic and prognostic markers involved in childhood ITP 10.3389/fmolb.2022.1007347 | 12-10-2022 | IFNG-AS1,GAS5,childhood ITP,qRT-PCR,diagnosis,prognosis,biomarkers | Background/aim: IFNG-AS1 is a long noncoding RNA that works as an enhancer for the Interferon-gamma (IFN-γ) transcript. GAS5 (growth arrest-specific 5) is a lncRNA that is associated with glucocorticoid resistance. Aberrant expressions of IFNG-AS1 and GAS5 are directly linked to numerous autoimmune disorders but their levels in childhood ITP are still obscure. This study aims to elucidate expressions of target lncRNAs in childhood ITP and their association with pathophysiology and clinical features of the disease as well as their association with types and treatment responses. Method: The fold changes of target lncRNAs in blood samples from children with ITP and healthy controls were analyzed using quantitative real-time PCR (qRT-PCR). Results: There were overexpressed lncRNAs IFNG-AS1 and GAS5 in serum of childhood ITP patients [(median (IQR) = 3.08 (0.2–22.39) and 4.19 (0.9–16.91) respectively, Also, significant higher IFNG-AS1 and GAS5 (p < 0.05) were present in persistent ITP (3–12 months) [ median (IQR) = 4.58 (0.31–22.39) and 3.77 (0.87–12.36) respectively] or chronic ITP (>12 months) [ median (IQR) = 5.6 (0.25–12.59) and 5.61 (1.15–16.91) respectively] when compared to newly diagnosed <3 months patients [IFNG-AS1 median (IQR) = 1.21 (0.2–8.95), and GAS5 median (IQR) = 1.07 (0.09–3.55)]. Also, significant higher lncRNAs IFNG-AS1 and GAS5 were present in patients with partial response to treatment [IFNG-AS1 median (IQR) = 4.15 (0.94–19.25), and GAS5 (median (IQR) = 4.25 (0.81–16.91)] or non-response [IFNG-AS1 median (IQR) = 4.19 (1.25–22.39) and GAS5 median (IQR) = 5.11 (2.34–15.27)] when compared to patients who completely responded to treatment (IFNG-AS1 median (IQR) = 2.09 (0.2–14.58) and GAS5 (median (IQR) = 2.51 (0.09–10.33). In addition, following therapy, the expressions of IFNG-AS1 and GAS5 are significantly negatively correlated with platelet count. Conclusion: Findings suggest that lncRNAs IFNG-AS1 and GAS5 are novel diagnostic and prognostic genetic markers for childhood ITP that can aid in a precise prediction of the disease’s progress at the time of diagnosis and could be a useful tool for treatment planning. | The Ifng antisense RNA 1 (IFNG-AS1) and growth arrest-specific transcript 5 (GAS5) are novel diagnostic and prognostic markers involved in childhood ITP 10.3389/fmolb.2022.1007347
Background/aim: IFNG-AS1 is a long noncoding RNA that works as an enhancer for the Interferon-gamma (IFN-γ) transcript. GAS5 (growth arrest-specific 5) is a lncRNA that is associated with glucocorticoid resistance. Aberrant expressions of IFNG-AS1 and GAS5 are directly linked to numerous autoimmune disorders but their levels in childhood ITP are still obscure. This study aims to elucidate expressions of target lncRNAs in childhood ITP and their association with pathophysiology and clinical features of the disease as well as their association with types and treatment responses. Method: The fold changes of target lncRNAs in blood samples from children with ITP and healthy controls were analyzed using quantitative real-time PCR (qRT-PCR). Results: There were overexpressed lncRNAs IFNG-AS1 and GAS5 in serum of childhood ITP patients [(median (IQR) = 3.08 (0.2–22.39) and 4.19 (0.9–16.91) respectively, Also, significant higher IFNG-AS1 and GAS5 (p < 0.05) were present in persistent ITP (3–12 months) [ median (IQR) = 4.58 (0.31–22.39) and 3.77 (0.87–12.36) respectively] or chronic ITP (>12 months) [ median (IQR) = 5.6 (0.25–12.59) and 5.61 (1.15–16.91) respectively] when compared to newly diagnosed <3 months patients [IFNG-AS1 median (IQR) = 1.21 (0.2–8.95), and GAS5 median (IQR) = 1.07 (0.09–3.55)]. Also, significant higher lncRNAs IFNG-AS1 and GAS5 were present in patients with partial response to treatment [IFNG-AS1 median (IQR) = 4.15 (0.94–19.25), and GAS5 (median (IQR) = 4.25 (0.81–16.91)] or non-response [IFNG-AS1 median (IQR) = 4.19 (1.25–22.39) and GAS5 median (IQR) = 5.11 (2.34–15.27)] when compared to patients who completely responded to treatment (IFNG-AS1 median (IQR) = 2.09 (0.2–14.58) and GAS5 (median (IQR) = 2.51 (0.09–10.33). In addition, following therapy, the expressions of IFNG-AS1 and GAS5 are significantly negatively correlated with platelet count. Conclusion: Findings suggest that lncRNAs IFNG-AS1 and GAS5 are novel diagnostic and prognostic genetic markers for childhood ITP that can aid in a precise prediction of the disease’s progress at the time of diagnosis and could be a useful tool for treatment planning.
A vast number of non-coding transcripts were discovered during human genome sequencing, they were formerly believed to be non-functional, however recently, scientists discovered that they have regulatory functions on gene expression involving transcriptional, posttranscriptional, and translational levels. Long noncoding RNAs (lncRNAs) are one type of them which defines as RNA transcripts with more than 200 nucleotides that are not translated into protein (Palazzo and Lee, 2015). As more evidence accumulates, lncRNAs appear to have an essential function in immune control, suggesting that aberrant lncRNAs expressions may have played a part in the triggering of autoimmunity comes in a variety of forms (Gao et al., 2018). IFNG-AS1 (Ifng antisense RNA 1), also known as Tmevpg1 (Theiler’s murine encephalomyelitis virus persistence candidate gene 1), or NeST, (nettoie Salmonella pas Theiler’s), is affiliated as lncRNA transcript which gene is found on the DNA strand opposite the interferon-gamma (IFN- γ) coding gene so hence its name. The coexpression of lncRNA, IFNG-AS1 with T-bet is related to the active transcription of IFN- γ in Th1 cells. Th1 cells are an inflammatory fraction of CD4+ T cells that primarily generate interferon-gamma (IFN-γ), an inflammatory cytokine that protects against intracellular pathogens, and is related to delayed-type hypersensitivity, and autoimmune diseases including ITP (Li et al., 2016), (Collier et al., 2012a). GAS5 (growth arrest-specific 5) is a lncRNA that was first discovered as a cancer tumor suppressor gene. GAS5 expression has been found to be abnormal in numerous malignancies, and this gene has been linked to cell cycle inhibition and apoptosis (Ji et al., 2019). Recently, scientists discovered that GAS5 dysregulation has been implicated in the development of autoimmunity states; The most effective medications for treating autoimmune illnesses are glucocorticoids (GC). GAS5 interacts with the DNA binding domain of glucocorticoid receptors (GRs), and preventing GRs from connecting with DNA, preventing glucocorticoid activity which is a powerful immunosuppressive and hence contributes to the development of autoimmune disorders (Moharamoghli et al., 2019). Immune thrombocytopenic purpura (ITP) is a blood disorder characterized by a decreased platelet count below 100×109/L with normal hemoglobin and leucocytes, when there are no additional causes or conditions that could be linked to thrombocytopenia, the condition is called idiopathic or primary ITP. The fundamental immunologic abnormality in ITP is thought to be humoral and cell-mediated antiplatelet antibodies thus leading to enhanced platelet destruction and often inappropriate platelet manufacture in the bone marrow. These phenotypic changes reflected a change in gene regulation (Makis et al., 2017), (Ayoub et al., 2020). Childhood Immune Thrombocytopenia (ITP) is one of the most frequent autoimmune hemorrhagic diseases affecting 5 to 10 children out of every 100,000. Twenty to thirty percent of cases were shown to be prone to chronicity with increased susceptibility to bleeding, infection, and blood malignancy than healthy children (Beshir et al., 2021) (Li et al., 2020). Previous studies demonstrated that when compared to children who received therapeutic regimens, the no therapy group had a significantly higher recurrence and more serious complications (Makis et al., 2017) (Kühne et al., 2011). The lncRNAs involved in immune mechanisms that play role in the initiation of the disease are the targets of therapy when indicated, Aberrant IFNG-AS1 and GAS5 expressions are linked to numerous autoimmune disorders suggesting the possible role of these genes in the pathogenesis of autoimmunity state (Li et al., 2018). There are two types of ITP; Adult ITP and childhood ITP (Kühne et al., 2011), although, two separate studies have reported that levels of IFNG-AS1and GAS5 in peripheral blood of adult ITP patients (Li et al., 2016), (Li et al., 2020), the expressions of IFNG-AS1 and GAS5 in childhood ITP are still unknown. Thus, we designated this study to investigate the expressions of lncRNAs IFNG-AS1 and GAS5 in childhood ITP and their relationship to clinical characteristics of the disease, as well as types and treatments.
Materials used to execute this work include; one- MiRNeasy Serum/Plasma extraction kit (Qiagen, Valencia, CA, USA) was used for total RNA extraction from serum. Two- The RT2 First Strand Kit (Qiagen, Valencia, CA, USA) was used for reverse transcription to produce cDNA. Three- The RT2 SYBR Green PCR kit (Qiagen, Maryland, USA) was used to perform qRT-PCR along with specific primers supplied by Qiagen, (Valencia, CA, USA); for IFNG-AS1 (Catalog no; 330701LPH20079A, Accession no, ENST00000536914.0) and GAS5 (Catalog no; 330701LPH11340A, Accession no, NR_002578.2) and we used GAPDH primer (Catalog no: 330701 LPH31725A, Accession no: ENST00000496049.0) to standardize the expression pattern and quantify the target long non-coding RNAs.
This case-control study was comprised of eighty-eight children (1–13 years old) who suffered from primary ITP and eighty-eight healthy age and sex-matched children with no history of any disease. Cases were gathered from the Pediatric Departments’ inpatient and outpatient clinics at Fayoum University Hospital in Egypt over a period of 6 months between Jan 2022 and June 2022.
A detailed history was taken including age, sex, complaint at presentation, disease duration, presence of bleeding manifestations, history of splenectomy, presence of preceding febrile illness, recent vaccination, helicobacter pylori infection history, HCV and/or HBV viral infection. Causes of secondary ITP such as autoimmune rheumatological diseases and the previous treatment regimen. A systematic clinical examination is performed, with special attention paid to clinical symptoms of bleeding such as bruises or purplish spots on the skin, mucous membranes, or gums, and any concomitant splenomegaly. Laboratory tests including complete blood count (CBC) and bone marrow examination to exclude other diseases resulting in thrombocytopenia were done. The definite diagnosis was established as primary idiopathic thrombocytopenia depending on the presence of solitary low platelet count (<100×109/L) in a patient with multiple bruises and/or petechiae.
The ITP patients were categorized regarding the duration of illness, 42 (47.72%) were newly diagnosed <3 months, 27 (30.86%) with persistent ITP 3–12 months, and 19 (21.60%) had chronic >12 months ITP. Treatment options regimen includes corticosteroids, IVIG, immunosuppressive drugs, and Eltrombopag (detailed treatment regimens are presented in Table 2). Response to treatment; 38 (43.18%) of patients showed complete response to treatment (CR) which is defined as platelet count ≥100×109/L without clinically significant bleeding, 43 (48.87%) showed partial response to treatment (PR); which is a platelet count between 30 and 100×109/L or twice the baseline platelet count without significant bleeding, and 7 (7.95%) with no response (NR) in patients with the present spleen is platelet count <30×109/L in two different measurements or the increase in platelet is less than twice baseline count (Makis et al., 2017)., (Rodeghiero et al., 2009)
Children aged 1–13 years, of both sexes, with primary ITP were included in the study while patients with secondary causes for purpura, splenectomy, with concurrent viral, or autoimmune diseases were excluded from the study or with a history of receiving medications cause thrombocytopenia.
Before the sample was taken, the study protocols were thoroughly explained to the parents, and they were asked to sign an informed consent form. The Ethics Committee at the Faculty of Medicine in Fayoum University reviewed and approved this work protocol no (R210-89). This work was done per the Declaration of Helsinki’s ethical requirements.
With the help of skilled medical staff, we withdrew 5 mL of venous blood from each participant into a plain tube for serum preparation and RNA extraction. For serum separation, samples were left on a straight surface for nearly 15 min, supernatants were transferred to clean tubes and centrifuged at 4,000 xg for 10 min to isolate serum which was kept at -80°C until use.
Total RNA was extracted from the serum according to the manufacturer’s instructions using the miRNeasy Serum/Plasma extraction kit (Qiagen, Valencia, CA, USA). The first step is to clean the area of work, pipettes, and centrifuge with prepared ethanol 70%, second step is aimed at lysis of cells and release of nucleic acid so after wearing clean gloves we added 1,000 μl QIAzol lysis reagent to 200 μL sample and incubated the mix after shaking vigorously at room temperature for 5 min. Thirdly, to enhance phase separation so that RNA is purified from DNA and protein debris, we added 200 μl chloroform, vortexed, and put the samples in the freezer for 5 min to avoid RNA destruction then the samples were centrifuged at 12.000 xg for 15 min. Then, we collected the upper aqueous parts in new tubes and added 1.5% of its volume to 100% ethanol to promote RNA precipitation. Afterward, we used the mini spin column in a 2 ml collection tube accompanied RNeasy kits to promote solid phase separation of RNA by adding the samples into two subsequent sessions (750 μl) each with centrifugation at 8,000 g for 15 s, discarded the flow-through and transferred the spin column containing binding RNA to clean 2 ml collection tubes. We used the washing buffers RWT and RPF in the kit; each spin column received 700 μl of RWT buffer, then centrifuged at 8,000 xg for (Fouad et al., 2022a) s, the flow-through was discarded, and the column was reused for the following phase, we next pipetted a 500 μl buffer RPE to the spin column and centrifuged it at 8,000 xg for 15s, discarded the flow-through water, and repeated the previous process. We moved the spin column to a fresh collecting tube and centrifuged it at full speed for 2 min to dehydrate RNA samples. Finally, for elusion, the spin column was transferred to a clean Eppendorf tube and 50 ul Rnase-free water was pipetted directly onto the column before centrifugation at 8,000 xg for 1 min. We next used a NanoDrop 1,000 spectrophotometer (Thermo Scientific, Waltham, MA, USA) to measure the RNA concentration and purification at 260/280 nm A ratio of 2.0 was considered “pure.”
The RT2 First Strand Kit (Qiagen, Valencia, CA, USA) was used for reverse transcription to produce cDNA, and the manufacturer’s procedure was followed. RT reaction was done in a final volume of 20 μl (10 μl reverse-transcription mix was added to each tube containing genomic DNA elimination mix), Conventional PCR was used to incubate the samples for 60 min at 37°C. followed by incubation for 5 min at 95°C to inactivate reverse transcriptase.
It has previously been shown that lncRNAs were expressed in serum (Mohammed et al., 2022)– (Visconti et al., 2020). The RT2 SYBR Green PCR kit (Qiagen, Maryland, USA) was used to evaluate the expression of the lncRNAs IFNG-AS1 and GAS5 in serum using specific primers supplied by Qiagen, (Valencia, CA, USA); for IFNG-AS1 (Catalog no; 330701LPH20079A, Accession no, ENST00000536914.0) and GAS5 (Catalog no; 330701LPH11340A, Accession no, NR_002578.2) and we used GAPDH primer (Catalog no: 330701 LPH31725A, Accession no: ENST00000496049.0) to standardize the expression pattern and quantify the target long non-coding RNAs (Li et al., 2016), (Peng et al., 2021). The reaction mix was prepared in a nuclease-free tube according to the manufacturer’s protocol for a 25 μl per well reaction volume. The Rotor-gene Q Real-time PCR system (Qiagen, USA) was used to perform quantitative real-time PCR under the following conditions: 95°C for 10 min, then 45 cycles of 15s at 95 and 60°C for 1 min.
Melting curve tests were carried out after the PCR cycles were completed to confirm the particular production of the expected PCR result. The cycle threshold (Ct) value is the number of qPCR cycles essential to the fluorescent signal to cross a specific threshold. By deducting the Ct values of GAPDH from those of the target long non-coding RNAs, ΔCt was determined, and by subtracting the ΔCt of the control samples from the ΔCt of the cases samples, ΔΔCt was determined. Then we used the subsequent equation (2−ΔΔCt equation) (Livak and Schmittgen, 2001) to demonstrate the fold change of target genes relative to controls which were set as 1.
Data was presented by mean ± SD (Standard Deviation), number and percentage, median and interquartile range; (IQR). SPSS version 22 (SPSS Inc) was used for analyzing data. The mean, SD, median, and range were calculated for the quantitative data. One-way ANOVA test was used for normal distribution data analysis in more than two groups. When variables were not normally distributed, the Mann–Whitney-U test (2 groups) or Kruskal Wallis test (more than two groups) was used in comparing groups. Otherwise, the independent-T test was used. The significance of the qualitative data was detected by chi-square (χ2). Pearson correlation was done to explore the association between IFNG-AS1 and GAS5 and the clinical parameters, treatment, and response to treatment. The sensitivity and specificity of IFNG-AS1 and GAS5 each alone or combined regarding the discrimination between ITP cases and healthy control subjects the receiver operating characteristic (ROC) curve analysis was done. All the results of the tests were interpreted as significant by considering that p ≤ 0.05.
The sample size equals 88 cases, to adapt the cost of research without decreasing the validity of results we used the formula; Necessary sample size = to get the appropriate sample size involved in this research, confidence levels 95% were converted into Z scores, the standard deviation was 0.5, and the margin of error was 0.05).
A total of 88 ITP children, ranging in age from 1 to 13 years were included in this case-control study with mean ± SD = 4.9 ± 2.15 years and 49 (55.68%) patients were females and 88 healthy children with ages ranging from 2 to 12 years with mean ± SD = 5.1 ± 2.09 years and 45 (51.13%) were females. This study was constructed to see how effective the lncRNAs IFNG-AS1 and GAS5 were at distinguishing between children within ITP and those within the control group and to explore the link between their levels and types of childhood ITP as well as with treatment regimens and response. Results represented in (Table 1) showed that the studied groups were matched regarding gender (p = 0.254) and age (p = 0.172). Regards CBC picture, ITP patients had lower hemoglobin (HB) (mean ± SD = 9.9 ± 2.19 gm%), lower platelet count (PL) either before (mean ± SD = 9.33 ± 2.51× 103) or after treatment (mean ± SD = 129.14 ± 48.95 × 103) and higher absolute lymphocytic count (ALC) (mean ± SD = 2,509 ± 926/ml) than controls (mean ± SD = 11.01 ± 1.58 gm% for HB, 191.35 ± 30.15 × 103 for PL count, and 1815 ± 875/ml for ALC, p < 0.05). There was no significant difference existed between the two groups regards white blood cell count (WBC) (mean ± SD = 8,073 ± 1,352/ml for ITP patients and 7,958 ± 1,250/ml for controls, p = 0.095).
The patients were divided into groups based on the duration of illness; 42 (47.72%) were newly diagnosed <3 months, 27 (30.86%) with persistent ITP 3–12 months, and 19 (21.60%) had chronic >12 months ITP. Out of 88 patients; 5 (5.68%) with positive family history, 31 (35.23%) with history of bleeding, 59 (67.05%) with preceding febrile illness, 5 (5.68%) with helicobacter infection, and 5 (5.68%) with splenomegaly. All patients had no history of splenectomy, hepatitis B virus (HBV), or hepatitis C virus (HCV) infection. Treatment regimens and responses to treatment are shown in (Table 2).
Children with ITP had significantly higher IFNG-AS1 [(median (IQR) = 3.08 (0.2–22.39), mean ± SD = 18.37 ± 19.54] and GAS5 [(median (IQR) = 4.19 (0.9–16.91), mean ± SD = 6.17 ± 11.33)] expressions in serum than that of healthy controls (p < 0.001 each) (Figure 1).
Regards the association between these genes and clinical characteristics, treatment and response to treatment in cases, we demonstrated that significant higher IFNG-AS1 and GAS5 (p < 0.05) were present in persistent ITP (3–12 months) [ median (IQR) = 4.58 (0.31–22.39) for IFNG-AS1 and 3.77 (0.87–12.36) for GAS5] or chronic ITP (>12 months) [ median (IQR) = 5.6 (0.25–12.59) for IFNG-AS1 and 5.61 (1.15–16.91) for GAS5] when compared to newly diagnosed <3 months patients [IFNG-AS1 median (IQR) = 1.21 (0.2–8.95), and GAS5 median (IQR) = 1.07 (0.09–3.55)]. Also, significant higher lncRNAs IFNG-AS1 and GAS5 were present in patients with partial response to treatment [IFNG-AS1 median (IQR) = 4.15 (0.94–19.25), and GAS5 (median (IQR) = 4.25 (0.81–16.91)] or non-response [IFNG-AS1 median (IQR) = 4.19 (1.25–22.39) and GAS5 median (IQR) = 5.11 (2.34–15.27)] when compared to patients with complete response to treatment (IFNG-AS1 median (IQR) = 2.09 (0.2–14.58) and GAS5 (median (IQR) = 2.51 (0.09–10.33). As well, significant high IFNG-AS1 was detected in patients with positive family history (median (IQR) = 9.09 (2.54–22.39) than in those with negative family history (median (IQR) = 1.98 (0.2–8.04), p = 0.04), just significant p-value = 0.05 was detected between positive family history patients and negative family history patients regards GAS5. In the cases group, there were no significant variations in target gene expressions in terms of gender, splenomegaly, a history of previous febrile illness or bleeding, helicobacter infection, or treatment regimens (Table 3).
Strong significant positive correlation was detected between IFNG-AS1 and GAS5 (r = 0.778, p < 0.001) (Figure 2), while negative correlations were detected between both genes and platelet count after treatment (r = -0.452, p < 0.001 for IFNG-AS1 and r = -0.438, p < 0.001 for GAS5) (Figures 3, 4). There was no correlation between target genes and age, HB concentration, WBCs count, ALC, or PL count before treatment (Table 4).
Figure 5 and Table 5 are shown the ROC curves of lncRNAs IFNG-AS1 and GAS5 in ITP patients. The ROC curves of lncRNAs IFNG-AS1 and GAS5 in ITP patients are demonstrating the diagnostic utility of these markers as predictors in distinguishing between patients with ITP and controls. LncRNA IFNG-AS1; AUC (95% CI) = 0.900 (0.769–0.980), p < 0.001, cut off point 1.38, sensitivity 66.3%, specificity 97.8%, total accuracy, 82.05%. Lnc RNA GAS5, AUC (95% CI) = 0.803 (0.625–0.975), p = 0.001, cut off point 2.58, sensitivity 68.5%, specificity 95.2%, total accuracy 81.85% and their combined expression AUC = 1.000, p < 0.001, sensitivity and specificity 100.0%. After reviewing the previous literature to explore the functions of target genes in ITP pathogenesis, we designed (Figure 6) to show how these genes could be targets of therapy in childhood ITP.
A. We designed Figure 6 to show the proposed function of target genes IFNG-AS1 and GAS5 in childhood ITP pathogenesis and their potential usage as targets of therapy. IFNG-AS1 is selectively expressed in T helper1 (Th1) cells by the effect of Stat4 and T-bet transcription factors (Collier et al., 2012b), upregulated IFNG-AS1 promoting IFN-γ expression and secretion (Petermann et al., 2019). Excessive IFN-γ decreases CD4+CD25+FoxP3+ Treg cell numbers (the major cells that relate to the maintenance of auto-tolerance) (Zufferey et al., 2017). Normalization of Tregs is an effective therapy for ITP (Semple et al., 2010), So, Knockdown of IFNG-S1 may be a potential new therapy for ITP (Figure 6A). B. GAS5 is associated with the glucocorticoid receptors (GRs) through hindering binding of GRs to target genes’ glucocorticoid receptor elements (GREs), and repressed GRs transcriptional activity of endogenous glucocorticoid-responsive genes (Wapinski and Chang, 2011; Kino et al., 2010). Hence, the knockdown in GAS5 levels may have an impact on the increased sensitivity of immunosuppressive treatment with glucocorticoid-related drugs (Figure 6B).
Childhood Immune Thrombocytopenia (ITP) is typically a benign self-limiting condition that resolves in a few months. However, between 20 and 30% of patients develop chronic ITP, which puts them at risk of serious complications including bleeding such as cerebral hemorrhage and blood cancer (Ayoub et al., 2020). Efficient treatment could decrease the susceptibility of ITP children to these serious complications (Makis et al., 2017) (Kühne et al., 2011). LncRNAs tangled in immune system regulation could be therapeutic targets (Zou and Xu, 2020). Abnormal IFNG-AS1 and GAS5 expression levels have been linked to a number of autoimmune disorders, implying that these genes may play a role in the pathogenesis of autoimmunity and could represent new targets of therapy (Wu et al., 2017), (Peng et al., 2020a). From the facts that; primary ITP provocation is a result of immune system gene dysregulation, lncRNAs are immune gene regulatory molecules, IFNG-AS1 and GAS5 are two lncRNAs that are directly linked to autoimmune disorders, previously related to adult ITP and no other previous study tested the expression of these genes in childhood ITP, we construct this case-control study to elucidate expressions of target lncRNAs in childhood ITP and their association with pathophysiology and clinical features of the disease as well as their association with types and treatment. Our results revealed that the age (mean ± SD) of our patients was 5.1 ± 2·19 years, and there is no sex prediction unlike in adults, the initial platelet count was 9.33 ± 2.51×103/L. these results agree with other studies for age, sex, and initial platelet counts (Talaat et al., 2014), (Diab et al., 2021). In this paper the common clinical presentation was bleeding ∼35% followed by splenomegaly ∼6%, in contrast, a study by Ayoub et al, 2020 showed rates of ∼ 27% for bleeding and ∼ 3% for splenomegaly (Ayoub et al., 2020). We reported that more than 50% of our patients had persistent and chronic ITP while prior studies documented a range between 30% and 40% of patients prone to the persistence of thrombocytopenia after 6 months of diagnosis (Makis et al., 2017), (Ayoub et al., 2020). It should be stressed that the etiology of ITP is unknown, and on the question of possible precipitating factors, this study found that nearly 6% of our patients have a positive family history, this was consistent with Ayoub et al, 2020 (Ayoub et al., 2020) but less than percentage documented by Diab et al, 2021 which reported 4% of patients had positive family history (Diab et al., 2021). History of helicobacter pylori infection accounted for 6% of our cases; other study reported 10% (Ayoub et al., 2020). While nearly 60% experience a period of febrile illness before disease, this was constant with other studies documented that two-thirds of patients had a preceding infection and febrile illness two to 3 weeks before the disease (Ayoub et al., 2020) (Livak and Schmittgen, 2001). The most important clinically relevant findings were overexpressed lncRNAs IFNG-AS1 and GAS5 in the serum of childhood ITP patients [(median (IQR) = 3.08 (0.2–22.39), mean ± SD = 18.37 ± 19.54 for IFNG-AS1, and median (IQR) = 4.19 (0.9–16.91) for GAS5 than controls, Also, significant higher IFNG-AS1 and GAS5 were linked to patients with positive family history, patients with persistent or chronic ITP than those who are newly diagnosed, patients with no response or partial response to treatment when compared with patients who completely responded to treatment. In addition, IFNG-AS1 and GAS5 expression were significantly negatively correlated with platelet count after therapy. Regards IFNG-AS1, our results could be explained by the findings of Collier et al, 2012 who implied that IFNG-AS1 is a part of a larger family of lncRNAs called lincRNAs that positively control gene transcription. They defined IFNG-AS1 as a Th1-specific lincRNA that promotes the transcription and secretion of the IFN-γ (Collier et al., 2012a). IFN-γ expression abnormalities have been linked to a variety of autoimmune disorders. In ITP patients, plasma IFN-γ levels were shown to be higher, which hastens disease development by lowering CD4+CD25+FoxP3+ Treg levels (responsible for preserving self-tolerance by interaction with APC and lowering CD19+ B cell and CD8+ T cell responses) (Zufferey et al., 2017), (Talaat et al., 2014). Many pieces of research have demonstrated that ITP patients have decreased Tregs, and interestingly, several medications that boost platelet cell number in ITP, such as glucocorticoid, rituximab, and intravenous immunoglobulin (IVIG), appear to do so via normalization of Tregs (Zufferey et al., 2017), (Semple et al., 2010), Thus these findings suggesting that silencing of IFNG-AS1 is a potential target of therapy for ITP. The previously mentioned data explain the significantly higher IFNG-AS1 levels presented in the partial response or no response patients group and also explain the significant negative correlation between IFNG-AS1 Levels and platelet number after therapy. A central contribution to recent work is a study that examined the levels of IFNG-AS1 in adult thrombocytopenic patients and found the transcript level of IFNG-AS1 in peripheral blood mononuclear cells (PBMCs) from active adult ITP patients was lower than in healthy controls. Furthermore, researchers discovered a positive correlation between IFN-γ and IFNG-AS1 transcript levels in PBMCs from healthy controls, as well as a similar change tendency after short-term activation. As a result, researchers speculated that IFNG-AS1 has been shown to enhance IFN-γ transcription. Furthermore, IFN-γ overexpression had a negative feedback effect on IFNG-AS1 expression, resulting in lower IFNG-AS1 levels in adult ITP patients (Li et al., 2016). Our findings are further supported by studies that investigated IFNG-AS1 in other autoimmune diseases which revealed that IFNG-AS1 expression from PBMCs was elevated in Hashimoto’s thyroiditis patients and contributed to Th1 cell response and IFN-γ expression levels which are implicated in the disease’s etiology (Peng et al., 2015). Likewise, lncRNA signatures in ulcerative colitis colonic tissues revealed increased expression of INFG-AS1 which is function as an enhancer of inflammation through positive regulation of IFN-γ expression. (Padua et al., 2016). In the same way, the transcript level of lncRNA IFNG-AS1 and its target gene IFN-γ were shown to be higher in the peripheral blood of rheumatoid arthritis patients (RA) than in controls and they were positively linked (Peng et al., 2020b). Additionally, Fouad et al, 2022 revealed that the expression level of IFNG-AS1 was upregulated significantly in the serum of patients with Behçet disease compared with controls (Fouad et al., 2022b). Regarding GAS5, the widely accepted function of GAS5 is that it is a tumor-suppressive lncRNA, which is implicated in a wide range of malignancies. Recently, scientists examined levels of GAS5 in immune-related diseases and elucidated its underlying molecular function, they concluded that the expression level of GAS5 was aberrant in patients with autoimmune disorders when compared to controls, and in addition to its tumor arresting function, GAS5 is a potent repressor of the glucocorticoid receptor (GR) (Mayama et al., 2016). Glucocorticoids (GCs) are a powerful immunosuppressant and are often the essential treatment of inflammatory and autoimmune illnesses, along with preventing rejection in transplanted patients. Several studies indicated that GAS5 is associated with glucocorticoid resistance through its direct attachment to the GR protein by competing with glucocorticoid receptor element (GRE) and acting as a decoy GRE, preventing glucocorticoid-induced gene transcription upregulation, and decreasing GCs activity so contributing to the development of numerous autoimmune diseases (Wu et al., 2020) (Suo et al., 2018). Corticosteroids, which are therapeutically variants of the glucocorticoids, are the principal therapy in ITP patients especially in chronic and resistant cases by different mechanisms include; raising the number of circulatory Tregs, reviving the Th1/Th2 proportion, restoring the Th17 count, constant with a rise in IL-10 and TGF- β, also, it alters B cell activation by lowering beta-cell stimulator (BlyS) and modulates dendritic cells (DCs) (Zufferey et al., 2017). Hence, the knockdown in GAS5 levels may have an impact on increasing the sensitivity of ITP patients to immunosuppressive treatment with glucocorticoid-related drugs. According to these data, we can interpret the significantly higher GAS5 associated with partial or no response patients group, and the negative correlation between GAS5 expression and platelet number after treatment. Support for this interpretation comes from Moharamoghli et al, 2019 who found that T cells from RA patients had higher amounts of GAS5 than those from controls (Moharamoghli et al., 2019), Also, Suo et al, 2018 documented that GAS5 and miR21 levels were considerably higher in CD4+ T cells from SLE patients than in control subjects, and GAS5 expression in CD4+ T cells was higher in ulcerated SLE patients than in non-ulcerated SLE patients (Suo et al., 2018). In a similar vein, Gharesouran et al, 2018 found that GAS5 levels were up-regulated in multiple sclerosis patients (Gharesouran, Taheri, Sayad, Ghafouri-Fard, Mazdeh, Davood Omrani). On the other hand, a recent study documented that GAS5 expression was downregulated in ITP patients’ PBMCs and ITP mice’s spleen tissues, and overexpression of GAS5 suppresses Th17 differentiation in vitro and relieved ITP in vivo via STAT3 reduction. Similarly, Li et al, 2019 concluded that the level of GAS5 was reduced in RA synovial tissues (Li et al., 2020). Furthermore, GAS5 and IL-10 mRNA levels in myasthenia gravis patients’ peripheral blood mononuclear cells (PBMCs) were significantly lower than in healthy controls (Peng and Huang, 2022). This inconsistency of results may be due to variances in samples and disease nature. A strong positive correlation between both genes was detected that confirms their synergistic effects. The ROC curve analysis for lncRNAs IFNG-AS1 and GAS5 in ITP patients are, demonstrating the diagnostic utility of these markers as predictors in distinguishing between patients with ITP and controls with sensitivity and specificity reached 100.0% in the case of the combination of both genes. From the preceding discussion, we speculated that both genes are diagnostic biomarkers for childhood ITP, and the increasing IFNS-AS1 and GAS5 expressions in childhood ITP patients may participate in the progress of the disease, especially its higher levels were significantly associated with persistent and chronic ITP patients than newly diagnosed patients. This finding agreed with Li et al, 2016 who documented associated high IFNG-AS1 with active ITP when compared to inactive ITP patients (Li et al., 2016) Also, high IFNG-AS1 was present in active ulcerative colitis (Padua et al., 2016) and high GAS5 was linked to active SLE (Suo et al., 2018). Thus, IFNG-AS1 and GAS5 may be used as predictors of the course of the disease. This study confirms that IFNG-AS1 and GAS5 may be potential targets of therapy as we found that higher target genes were detected in patients with no or partial response to treatment when compared to patients who completely responded to treatment, these results further support the idea of GAS5 function as glucocorticoid resistant enhancer (Mayama et al., 2016). In addition, we conveyed that IFNG-AS1 and GAS5 expressions were significantly negatively correlated with platelet count after therapy, this result was consistent with Li et al, 2016 who documented a negative correlation of IFNG-AS1 with platelet count, but they found no correlation with treatment or duration of the disease (Li et al., 2016). The limitations of the current study include; 1) The sample size of the current study was modest and selected from the same geographical area which provoked the possibility of selection bias. 2) Insufficiency of scientific literature demonstrating the precise function of target lncRNAs in general and in particular to pediatric diseases including childhood ITP. Further multicentric studies are needed to confirm our findings, verify the utility of target genes as targets of therapy, and better understand the molecular role of IFNG-AS1 and GAS5 in the etiology and pathogenesis of childhood ITP. The current study contains several new and important insights; findings suggest that lncRNAs IFNG-AS1 and GAS5 are novel diagnostic and prognostic biomarkers for childhood ITP that can aid in a precise prediction of the disease’s progress at the time of diagnosis and could be a useful tool for treatment planning, reducing the risk of bleeding while avoiding drug side effects. We speculated that both genes have underlying molecular contributing roles in the development and prognosis of childhood ITP and may be used as targets of therapy to reduce the propagation of the disease to a chronic state.
LncRNAs IFNG-AS1 and GAS5 are novel diagnostic and prognostic genetic markers for childhood ITP. | true | true | true |
|
PMC9597442 | Bich Phuong Bui,Phuong Linh Nguyen,Ha Thi Thu Do,Jungsook Cho | Anxiolytic effect of Korean Red Ginseng through upregulation of serotonin and GABA transmission and BDNF expression in immobilized mice | 05-08-2022 | anxiolytic effect,GABA transmission,immobilization-induced anxiety-like behaviors,Korean Red Ginseng,serotonin transmission | Background Anxiolytic properties of Korean Red Ginseng (KRG) have been previously reported. However, the exact mechanism(s) of action remains to be elucidated. The present study investigated the effect of KRG on immobilization-induced anxiety-like behaviors in mice and explored the involvement of the serotonin and GABA systems and BDNF in the anxiolytic action. Methods Mice were orally administered with KRG (200 mg/kg/day) for 4 weeks and immobilized once daily for 2 h. p-Chlorophenylalanine (p-CPA) was intraperitoneally injected on day 22-28, and flumazenil or bicuculline was injected on day 25-28. After behavioral evaluations, brains were dissected for biochemical analyses. Results KRG improved immobilization-induced anxiety-like behaviors in mice, as assessed by the elevated plus maze (EPM) and marble burying tests (MBT). The anxiolytic effect of KRG was comparable to that of fluoxetine, a reference drug clinically used for anxiety disorders. A serotonin synthesis inhibitor, p-CPA, blocked the effect of KRG in the EPM and MBT, indicating the requirement of serotonin synthesis for anxiolytic action. In addition, the anxiolytic effect of KRG was inhibited by bicuculline (a GABAA antagonist) in MBT, implying the involvement of GABA transmission. Western blotting analyses revealed that KRG upregulated the expression of tryptophan hydroxylase and GABAA receptor in the brain, which was blocked by p-CPA. Enhanced BDNF expression by KRG in the hippocampus was also indicated to mediate the anxiolytic action of KRG in immobilized mice. Conclusion KRG exhibited the anxiolytic effect in immobilized mice by multiple mechanisms of action, involving enhanced serotonin and GABA transmissions and BDNF expression. | Anxiolytic effect of Korean Red Ginseng through upregulation of serotonin and GABA transmission and BDNF expression in immobilized mice
Anxiolytic properties of Korean Red Ginseng (KRG) have been previously reported. However, the exact mechanism(s) of action remains to be elucidated. The present study investigated the effect of KRG on immobilization-induced anxiety-like behaviors in mice and explored the involvement of the serotonin and GABA systems and BDNF in the anxiolytic action.
Mice were orally administered with KRG (200 mg/kg/day) for 4 weeks and immobilized once daily for 2 h. p-Chlorophenylalanine (p-CPA) was intraperitoneally injected on day 22-28, and flumazenil or bicuculline was injected on day 25-28. After behavioral evaluations, brains were dissected for biochemical analyses.
KRG improved immobilization-induced anxiety-like behaviors in mice, as assessed by the elevated plus maze (EPM) and marble burying tests (MBT). The anxiolytic effect of KRG was comparable to that of fluoxetine, a reference drug clinically used for anxiety disorders. A serotonin synthesis inhibitor, p-CPA, blocked the effect of KRG in the EPM and MBT, indicating the requirement of serotonin synthesis for anxiolytic action. In addition, the anxiolytic effect of KRG was inhibited by bicuculline (a GABAA antagonist) in MBT, implying the involvement of GABA transmission. Western blotting analyses revealed that KRG upregulated the expression of tryptophan hydroxylase and GABAA receptor in the brain, which was blocked by p-CPA. Enhanced BDNF expression by KRG in the hippocampus was also indicated to mediate the anxiolytic action of KRG in immobilized mice.
KRG exhibited the anxiolytic effect in immobilized mice by multiple mechanisms of action, involving enhanced serotonin and GABA transmissions and BDNF expression.
Anxiety disorders are the most common type of psychiatric disorders. According to large population-based surveys, anxiety affects up to 33.7% of the population during their lifetime, becoming an increasing psychoclinical challenge worldwide [1]. The current conceptualization of the etiology of anxiety includes an initial exposure to unavoidable factors like chronic stress, trauma, or a genetic vulnerability, resulting in neurobiological and neuropsychological alterations [2,3]. Despite the effectiveness of current pharmacotherapies for anxiety disorders, treatment failures may occur due to delayed responsiveness or unresponsiveness to drugs or intolerable adverse effects [4,5]. To overcome these challenges, there has been growing attention toward alternative therapeutic strategies, particularly targeting natural products or herbal medicine [6,7]. Korean Red Ginseng (KRG), produced from the roots of Korean ginseng (Panax ginseng Meyer, Araliaceae) cultivated for 4─6 years through repeated steaming and drying processes, has been widely used in traditional medicine for numerous neuropsychiatric disorders, including anxiety and depression [6,7]. However, only a few studies have been reported to show the anxiolytic effects of KRG in mice or rats. Red varieties of P. ginseng was shown to exert anxiolytic action in the open field test and elevated plus maze (EPM) test, and this effect was comparable to diazepam [8]. In another study, while KRG water extract did not increase the percentage of open arm entries or the time spent in open arms in the EPM test, a single oral administration of KRG butanol fraction was shown to exhibit anxiolytic-like effects in mice [9]. Moreover, KRG attenuated anxiety-like behaviors in rats during ethanol withdrawal [10]. Although few studies have identified the active constituent(s) of KRG, ginseng saponins are suggested to play important roles in anxiolytic actions [9]. For example, the crude saponin fraction and several ginsenosides from KRG, such as Rg3 and Rh2, exhibited anxiolytic effects in the EPM model [11]. Interestingly, the anxiolytic-like activities of Rg3 and Rh2 were antagonized by flumazenil (a benzodiazepine antagonist), suggesting their actions via the GABA/benzodiazepine system. GABA is a well-characterized neurotransmitter associated with anxiety disorders [12]. The lack of inhibitory neurotransmitter GABA and its major receptor (GABAA receptors) plays important roles in the pathophysiology of anxiety, which provides pharmacological rationale to treat anxiety disorders with benzodiazepines, such as diazepam [12]. However, several associated issues, including drug dependence, rebound anxiety, and memory impairment, restrict their use to short-term treatment of acute anxiety [13]. In addition to GABA, serotonergic dysfunction has also been associated with the etiology of anxiety disorders [4,14]. Due to the favorable benefit-risk ratio, selective serotonin reuptake inhibitors, such as fluoxetine (FLX), and serotonin-norepinephrine reuptake inhibitors are recommended as first-line drugs to treat anxiety [4,15]. However, the onset of the anxiolytic effect of these drugs has a latency of 2-4 weeks after administration. Although the involvement of serotonergic system in the antidepressant effect of KRG or ginsenosides is well-characterized [16,17], its role in the anxiolytic effect is not elucidated yet. This study aims to characterize the anxiolytic effect of KRG in a mouse model of anxiety induced by immobilization stress, using the EPM and marble burying tests (MBT). To elucidate the role of serotonin system, we investigate the impact of p-chlorophenylalanine (p-CPA, a selective 5-HT synthesis inhibitor) on the anxiolytic effect of KRG in mice. Moreover, since the modulation of GABA system has been proposed to mediate anxiolytic action of KRG [11], we also examine the effects of flumazenil and bicuculline (a competitive GABAA antagonist). Furthermore, biochemical changes mediating the serotonin and GABA systems are studied in the mouse brain. FLX is used as a positive reference drug in this study.
The standardized KRG extract was manufactured from the roots of 6-year Korean ginseng by the Central Research Institute, Korea Ginseng Corporation (Daejeon, Korea), and kindly provided (Lot No. H1312-9040). According to the manufacturer's data, the main components of KRG are: ginsenosides Re (0.82 mg/g), Rf (1.37 mg/g), (S)-Rg2 (1.50 mg/g), Rb1 (5.85 mg/g), Rc (2.29 mg/g), Rb2 (2.17 mg/g), Rd (0.89 mg/g), (S)-Rg3 (4.43 mg/g), (R)-Rg3 (2.02 mg/g), and Rh1 (1.28 mg/g). KRG was dissolved in distilled water to get the working solution of 20 mg/mL. FLX, flumazenil, and bicuculline were bought from Sigma-Aldrich (St. Louis, MO, USA), and p-CPA was purchased from Tocris (Bristol, UK). All other chemicals were of analytical grade.
Male ICR mice (20─25 g) were obtained from Orient Bio (Gyeonggi, Korea). The animals were maintained under a controlled temperature (22 ± 2°C) and relative humidity (40─60%) with a 12-h light-dark cycle, and given free access to a standard chow diet and water. All experimental procedures including the use, care, and handling of animals were conducted following the international guidelines (Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources, Commission on Life Sciences, National Research Council; National Academy Press: Washington D.C., 1996). Prior to the study, the rationale, design, and protocols of the experiments were approved by the Institutional Animal Ethical Committee of Dongguk University (approval number: IACUC-2019-001-2). The animals were randomly allocated into seven groups of nine mice each: Vehicle-treated control group; Vehicle-treated immobilized group; FLX-treated immobilized group; KRG-treated immobilized group; KRG + p-CPA-treated immobilized group; KRG + flumazenil-treated immobilized group; and KRG + bicuculline-treated immobilized group. All animals were subjected to immobilization stress during day 1─27, except for the vehicle-treated control group. Experimental drugs were administered to each group as follows: To the FLX group, FLX (10 mg/kg) were orally administered for 4 weeks. To the KRG-treated groups (KRG, KRG + p-CPA, KRG + flumazenil, and KRG + bicuculline), KRG (200 mg/kg) was orally administered for 4 weeks. The dosage of KRG was determined based on the previous reports [[18], [19], [20]]. As depicted in Fig. 1, p-CPA (100 mg/kg) was intraperitoneally injected on day 22─28, and flumazenil (3 mg/kg) or bicuculline (0.7 mg/kg) was injected on day 25─28, 15 min before KRG administration. The dosages, routes of administration, and duration of FLX and p-CPA [16], flumazenil [21], and bicuculline [22] were chosen from previous studies. Corresponding volumes of vehicle were given to the animals in control and vehicle-treated immobilized groups, instead. To induce chronic anxiety-like behaviors, the mice were subjected to immobilization stress according to the previous reports [23,24]. Briefly, 1 to 2 h after KRG or FLX administration, mice in the immobilized groups were restrained in 50-mL conical tubes (3 cm in diameter and 10 cm in length) once daily for 2 h. A breathing hole (0.3 cm in diameter) was inserted at the end of each tube to allow air to pass directly into the nose of the mouse. Body weights were measured daily for 4 weeks. Behavioral tests, including the rota-rod test (RRT), EPM, and MBT, were then performed on day 27─28 (Fig. 1), and analyzed by investigators blinded to the treatment conditions. Following the behavioral tests, all animals were sacrificed and their brains were immediately dissected for western blotting and immunohistochemistry (IHC). The experimental schedules are collectively summarized in Fig. 1.
RRT was performed to assess the motor coordination of animals in each group, using the rota-rod apparatus for mice (Ugo Basile Corporation, Varese, Italy), as described previously with minor modifications [25]. It was set with an accelerated velocity from 4 to 40 rpm. On the day before the test session (day 27), mice were trained to the rotating rods for 5 min at 1 h after drug administration. During the test session, endurance time (s) and the number of falls (falling frequency) were recorded for 5 min.
The EPM test was performed according to the procedures reported previously with minor modifications [9,26]. The EPM apparatus is comprised of two closed arms (74 cm × 6 cm) enclosed by walls with a height of 20 cm and two open arms (74 cm × 6 cm) without walls. These arms are connected by a central (6 cm × 6 cm) square. The entire maze was elevated 50 cm above the floor. Animals were placed on the middle square facing a closed arm and allowed to explore the apparatus for 10 min. The movement of animals was recorded using a video camera-based EthoVision Maze Test System (Noldus Information Technology, Wageningen, Netherlands). Time spent in open arms or closed arms (s), number of open arm entries, and total moved distance (cm) were determined.
The MBT was performed as previously described with some modifications [27]. Briefly, clean rat cages (17 cm × 25 cm × 40 cm) were filled with wood chip bedding to a depth of 4 cm and evenly spread to a flat surface. Sixteen glass marbles were then placed at equal intervals on the bedding in a 4 × 4 grid. Mice were placed in the middle of the cage and allowed to bury marbles for 30 min without disturbance. The marbles covered with bedding up to 2/3 of the surface area were counted.
After behavior tests, animals were sacrificed and their brains were quickly removed. The brains were dissected to obtain the cortex and the hippocampus, which were then separately homogenized in cold lysis buffer, as previously described [28]. The homogenates were centrifuged at 14,000 rpm for 30 min at 4°C, and the supernatants containing extracted proteins were collected and stored at ─80°C until used for biochemical study. Electrophoresis and immunoblotting were then performed according to procedures described previously [29], with primary antibodies specifically recognizing tryptophan hydroxylase (TPH), glutamic acid decarboxylase with a molecular weight of 67 kDa (GAD67) (GAD1, Cell Signaling Technology, Danvers, MA, USA), GABAA receptor, and brain-derived neurotrophic factor (BDNF) (Abcam, Cambridge, MA, USA). Blots were visualized with a Bio-Rad ChemiDoc XRS imaging system using enhanced chemiluminescence reagents (Bio-Rad, Hercules, CA, USA).
IHC was conducted as previously described with some minor modifications [28]. Briefly, mice were perfused with phosphate-buffered saline (PBS), followed by 4% paraformaldehyde. The brains were removed, fixed overnight, cryoprotected with 30% sucrose at 4°C, and embedded with optimum cutting temperature solution. The embedded brains were then cut with a cryostat (Leica Microsystems Ltd., Nuβloch, Germany) into 10-μm sections, mounted onto poly-L-lysine-coated slides, and surrounded by a hydrophobic barrier with a wax pen (ImmEdge Hydrophobic Barrier PAP Pen, Vector Laboratories, Burlingame, CA, USA). Non-specific binding was blocked with 5% goat serum and 0.5% Triton X-100 in PBS for 1 h at room temperature. The sections were incubated with anti-BDNF antibody (1:100 dilution) overnight at 4°C. After washing with PBS three times, the sections were incubated with fluorescent-conjugated secondary antibody (1:400 dilution) at room temperature for 1 h in the dark and mounted with anti-fade mounting medium containing DAPI. Fluorescence was visualized using a Nikon confocal laser-scanning microscope (Nikon Instruments Inc., Melville, NY, USA).
The quantitative data were presented as mean ± SEM. Statistical analysis was conducted by one-way ANOVA, followed by Tukey's post hoc test using SigmaPlot 12.5 software (Systat Software Inc., San Jose, CA, USA). A p < 0.05 was considered significant.
The body weights of mice were measured daily for 28 days. All animals gained body weight during 28 days of the experiment. However, the weight gain of animals in the immobilized groups was significantly less than that in the control group, starting from day 4. Neither KRG nor FLX restored the reduced rate of weight gain by immobilization. In addition, p-CPA, flumazenil, or bicuculline administration along with KRG did not significantly change the body weights of the immobilized mice (Fig. 2A).
Prior to investigating the effect of KRG on the immobilization-induced anxiety-like behaviors and its underlying mechanism(s), RRT was carried out to examine whether our experimental drugs, including KRG, could influence motor coordination in mice. The falling frequency in the control group appeared to be relative high, probably due to the short training period [30]. However, there were no significant changes in endurance time and falling frequency in all animal groups (Fig. 2B), indicating that our experimental drugs did not affect motor coordination and balance. Behavioral studies, including the EPM and MBT, were then conducted to evaluate the effect of KRG on the immobilization-induced anxiety-like behaviors.
The EPM test is commonly used to evaluate the effects of drugs on exploratory behaviors and anxiety in rodents. Increased open arm entries and time spent in the open arms indicate anxiolytic activities in the test [26,31]. To evaluate the anxiolytic effect of KRG, anxiety-like behaviors were induced in mice by immobilization stress, and EPM tests were performed after oral administration of KRG (200 mg/kg) for 4 weeks. In the EPM test, the time spent in open arms, the time spent in close arms, the number of open arm entries, and the total moved distance were determined (Fig. 3). The time spent in open arms was markedly reduced in the vehicle-treated immobilized group (Fig. 3A), indicating that immobilization successfully induced anxiety-like behaviors. While the number of entries to open arms in this group also showed a tendency to decrease (Fig. 3C), the time spent in close arms was slightly increased (Fig. 3B). However, the differences were not statistically significant from those in the non-immobilized control group (Fig. 3B and C). KRG markedly reversed the reduced time spent in open arms, the increased time spent in close arms, and the decreased number of open arm entries in the immobilized mice (Fig. 3A–C), demonstrating its anxiolytic effect. Similar effects were produced by FLX (10 mg/kg), a positive reference drug, confirming its anxiolytic effect in this study (Fig. 3A–C). By contrast, the total moved distance in the EPM test was not significantly altered in all experimental groups, including the vehicle-treated immobilized group (Fig. 3D). Thus, the induction of anxiety-like behaviors by immobilization stress was not attributed to the alterations in locomotor activities of mice. As shown in the representative tracking maps (Fig. 3E), the movement in open arms (red color) of vehicle-treated immobilized mice was dramatically reduced compared to that of the non-immobilized control group. In agreement with our findings shown in Fig. 3A–C, the immobilization-induced reduction of the movement in open arms was noticeably reversed by KRG or FLX (Fig. 3E). Collectively, these results demonstrated that KRG and FLX orally administered for 4 weeks have anxiolytic effects in the mice exposed to chronic immobilization stress.
To examine whether the anxiolytic effect of KRG was mediated by serotonergic and/or GABAergic transmissions, the KRG-treated immobilized mice were intraperitoneally injected with a serotonin synthesis inhibitor (p-CPA, 100 mg/kg), a benzodiazepine antagonist (flumazenil, 3 mg/kg), or a competitive GABAA receptor antagonist (bicuculline, 0.7 mg/kg), as illustrated in Fig. 1, and EPM tests were performed. The anxiolytic effect of KRG was notably inhibited by p-CPA, reducing the time spent in open arms and the number of open arm entries (Fig. 3A, C, and E). However, the administration of flumazenil or bicuculline did not inhibit the anxiolytic effect of KRG in this study. These results indicate that serotonin synthesis is required for the anxiolytic effect of KRG. By contrast, the anxiolytic action of KRG evaluated by the EPM test appears not to be mediated by GABA transmission.
To confirm our findings, we performed MBT, another behavioral test frequently used to measure anxiety-like behaviors in mice [27]. Consistent with the previous report [32], the vehicle-treated immobilized mice showed a marked increase in the number of buried marbles, an indicator of anxiety-like behaviors in MBT (Fig. 4A, black bar; Fig. 4B). This impact was dramatically reversed in mice by orally administered KRG for 4 weeks, demonstrating the anxiolytic effect of KRG in MBT. As expected, FLX also exhibited the anti-anxiety effect (Fig. 4A and B). The involvement of serotonergic and/or GABAergic transmissions in the anxiolytic action of KRG was also evaluated in MBT. Similar to the findings from the EPM test, p-CPA treatment completely blocked the anxiolytic effect of KRG in MBT (Fig. 4), revalidating the critical role of serotonin synthesis in the anxiolytic action. Interestingly, the bicuculline-treated group showed a remarkable increase in the number of buried marbles, suggesting that GABA transmission was also involved in the KRG effect. Although the number of buried marbles was also increased by flumazenil, this effect was not statistically significant (Fig. 4A).
Our behavioral studies using the EPM and MBT showed that p-CPA markedly abolished the anxiolytic effect of KRG (Fig. 3, Fig. 4), indicating the requirement of serotonin synthesis for its action. These findings prompted us to examine whether KRG altered the expression of TPH, an enzyme that catalyzes the initial and rate-limiting step in the synthesis of serotonin, in the brains of immobilized mice. We found that the levels of TPH in the cortex and hippocampus of the vehicle-treated immobilized mice were comparable to those of non-immobilized control mice (Fig. 5A and B). However, the TPH expression in the cortex was substantially increased by KRG administration, which was completely abolished by p-CPA, flumazenil, or bicuculline (Fig. 5A). Although both flumazenil and bicuculline failed to block the anxiolytic effect of KRG in the EPM test, bicuculline was found to significantly attenuate the KRG effect in MBT (Fig. 4A), suggesting potential involvement of GABA transmission. To test this possibility, we examined whether KRG altered the levels of GAD67 (the predominant isoform of GAD in the brain, catalyzing the decarboxylation of glutamate to synthesize GABA) and GABAA receptor in the brains of immobilized mice. While the level of GAD67 was not altered in the cortex of the vehicle-treated immobilized mice, its expression was reduced in the hippocampus compared to the control group (Fig. 5). Interestingly, both KRG and FLX increased the GAD67 expression in the cortex (Fig. 5A). However, it was not significantly altered in the hippocampus by any experimental drug tested in this study (Fig. 5B). The levels of GABAA expression were considerably downregulated by immobilization stress in the cortex and hippocampus. The reduced cortical GABAA receptor expression was completely restored by KRG or FLX (Fig. 5A). The KRG-induced restoration of GABAA was notably inhibited by p-CPA, but not by flumazenil or bicuculline. Collectively, these results indicate that enhanced serotonergic and GABAergic transmissions mediate the anxiolytic effect of KRG, mainly through upregulation of TPH and GABAA expression in the brain. Given that the upregulated TPH and GABAA receptor by KRG are distinctly inhibited by p-CPA, serotonin synthesis may be a crucial process to exert the anxiolytic action.
BDNF, a member of the neurotrophin family, plays an important role in stress-related mental disorders, such as anxiety and depression [33]. The BDNF expression is significantly affected by stress in specific brain regions. Accumulating evidence revealed that stress-associated anxiety-like phenotypes were closely related to the reduced BDNF level in the hippocampus [33]. Therefore, we evaluated the effects of KRG on the expression of BDNF in the hippocampus. In accordance with the previous reports, we also observed in western blotting analysis that the immobilization stress markedly decreased BDNF expression in the hippocampus, as noticed by the reduced intensity of band at 14 kDa corresponding to the mature form of BDNF (Fig. 6A). The reduced BDNF expression was notably upregulated by KRG or FLX. The KRG-induced upregulation of hippocampal BDNF was significantly reduced by p-CPA. Flumazenil or bicuculline also showed a tendency to attenuate the upregulated hippocampal BDNF, but the effect was not statistically significant. These findings were exactly reproduced in hippocampal slices by IHC analysis (Fig. 6B). Although the BDNF level in the cortex of immobilized mice was slightly decreased, the cortical BDNF expressions were not significantly altered by the experimental drugs tested in this study (Fig. 6A). Based on our findings, the upregulated BDNF expression by KRG in the hippocampus is also associated with the anxiolytic effect of KRG in immobilized mice.
Beneficial effects of P. ginseng and its constituents have been reported in various neuropsychological conditions, such as learning and memory deficits, mental illnesses, and cerebral ischemia [[34], [35], [36]]. In addition, the antidepressant and anti-anxiety effects of KRG and selected ginsenosides have been documented in depressive patients and various animal models [8,11,28,37]. Multiple mechanisms have been suggested to mediate their antidepressant activities, including increases in the levels of serotonin, dopamine, and norepinephrine, upregulation of BDNF, and modulation of the hypothalamic−pituitary−adrenal axis [38,39]. By contrast, the GABA/benzodiazepine system has been proposed to be the major mechanism mediating the anti-anxiety effects of KRG and selected ginsenosides, including Rg3, Rh2, and Rg1 [11,40,41]. Although several studies have emphasized involvement of the serotonergic system in the antidepressant effect of KRG [16,17], its role in the anxiolytic effect remains to be further delineated. The anxiolytic effects of KRG and ginsenosides have been mostly studied using naïve mice and rat models of post-traumatic stress disorder or ethanol withdrawal [8,10,[41], [42]]. The present study evaluated the anxiolytic effect of KRG using a mouse model of anxiety induced by chronic immobilization stress. In addition, to elucidate underlying action mechanisms(s) of KRG, we explored involvement of the serotonin and GABA systems and the BDNF expression in the anxiolytic action. Considering that current pharmacotherapy to treat anxiety disorders usually requires long-term treatment, and that it usually takes at least 2−4 weeks of drug administration to exert their efficacy [4,5], we designed our experimental protocol with the duration of KRG administration for 4 weeks in this study (Fig. 1). It is well-recognized that body weight changes are closely associated with anxiety [43]. Thus, prior to investigating the effect of KRG on immobilization-induced anxiety-like behaviors in mice, body weight changes were monitored for 28 days of our experiment. We found that immobilization stress reduced the rate of weight gain, which was not restored by the administration of any experimental drugs tested in this study (Fig. 2A). We also examined the motor coordination of animals in all groups by RRT on day 27, and found that our experimental drugs did not affect the motor coordination and balance (Fig. 2B). We then conducted behavioral studies to evaluate the anxiolytic effect of KRG on the immobilization-induced anxiety-like behaviors. Exploration-based models, including the EPM and MBT, are commonly utilized in a majority of rodent studies to characterize behavioral alterations in animals exposed to anxious stimuli, such as fear-induced avoidance [26,27,44]. In our study, immobilization stress-induced anxiety-like behaviors were evident from the marked reduction of time spent in open arms in the EPM test and the increased number of buried marbles in the MBT (Fig. 3, Fig. 4). Administration of KRG dramatically restored the time spent in open arms as well as closed arms and the number of open arm entries in the EPM test. Similarly, KRG effectively reversed the number of buried marbles in MBT, collectively demonstrating the anxiolytic effect of KRG. The effect of KRG was comparable to that of FLX (Fig. 3, Fig. 4), a well-known drug clinically used for the treatment of various types of anxiety disorders, such as panic disorder, generalized anxiety disorder, and obsessive-compulsive disorder [4,5]. Our findings ascertaining the anxiolytic effects of KRG and FLX in the immobilized mice are consistent with the previous findings examined in different animal models of anxiety [10,28]. One of the most widely acknowledged hallmarks associated with the pathophysiology of anxiety is the declines in neurotransmitters, such as GABA, serotonin, and dopamine [12,45], which may cause dysfunction of their receptors, resulting in impaired transmission in the brain. To establish involvement of the serotonin system in the anxiolytic effect of KRG, we used a serotonin synthesis inhibitor, p-CPA. As shown in Fig. 3, Fig. 4, we found that intraperitoneally injected p-CPA completely abolished the anxiolytic effect of KRG in the EMP and MBT, demonstrating the crucial role of serotonin synthesis. Serotonin is synthesized from tryptophan in two steps, with TPH as the rate-limiting enzyme. Thus, we then examined whether KRG altered the expression of TPH in the brains of immobilized mice. Upon KRG administration, confirmatory evidence of a substantially increased TPH level in the cortex was observed, and this upregulation was completely abolished by p-CPA (Fig. 5A). These results suggest that KRG enhances serotonin synthesis through upregulation of TPH, which contributes to its anxiolytic effect. To examine whether KRG increased serotonin levels in the brain, we measured the level of serotonin in the cortical homogenates using an ELISA kit (Abcam, Cat. No. ab133053). As expected, the serotonin level was considerably decreased in the brain of immobilized mice, and the decreased serotonin level was dramatically recovered by KRG or FLX to the level of control group (data not shown). Direct measurement of 5-HT levels in the mouse brain could strengthen this finding. Among the serotonin receptor subtypes identified, 5-HT2A and 5-HT1A receptors are reported to be most closely related to anxiety-like behavior in post-traumatic stress disorder in mice [46]. Thus, it would be interesting to identify subtypes of the receptor mediating the effect of KRG in immobilized mice. Although KRG has been reported to prevent depression-like behaviors in post-traumatic stress disorder through enhancing 5-HT concentration in the hippocampus [17], our study is the first to elucidate the requirement of serotonin synthesis for the anxiolytic action of KRG. Apart from the serotonin system, we also investigated the role of GABA transmission in the anxiolytic effect of KRG. While the administration of flumazenil or bicuculline did not counteract the anxiolytic effect of KRG in our EPM test (Fig. 3), the bicuculline-treated group showed a remarkable increase in the number of buried marbles in the MBT (Fig. 4), suggesting that the GABA system might be involved in the KRG effect. The GABA/benzodiazepine system has been previously reported to mediate anxiolytic-like activities of KRG and several ginsenosides, such as Rg3, Rh2, Rb1, and Rg1 in the EPM test [11,40,41]. The discrepancy between our results and previous findings in the EPM test might be, at least in part, due to the differences in animal models of anxiety. While we used the anxiety-like animal model induced by chronic immobilization stress, previous studies used naïve animals to test the anxiolytic effects [11,40,41]. An additional investigation should be conducted to further clarify the discrepancy. To examine the potential involvement of the GABA system, we further examined the effect of KRG on the expressions of GAD67 and GABAA in the brains of immobilized mice. We found that KRG and FLX increased the cortical GAD67 expression. However, the level of hippocampal GAD67 was not altered by KRG or FLX, compared to that of the control (Fig. 5). It remains to be determined whether the increased levels of GAD67 by KRG or FLX in the cortex play any role in their anxiolytic effects. Unlike GAD67, the levels of GABAA expression in the cortex and hippocampus were considerably downregulated by immobilization stress, and these levels were completely restored by KRG or FLX (Fig. 5). Collectively, our data demonstrate that KRG exerts its anxiolytic effect by enhancing serotonin transmission through upregulation of TPH and restoring the disrupted GABA transmission through upregulation of GABAA receptors in the immobilized mice. Interestingly, our findings imply that there may be mutual interactions between the serotonin and GABA systems to mediate the anxiolytic effect of KRG, as p-CPA and flumazenil or bicuculline reduced the expression of GABAA and TPH, respectively (Fig. 5). In support of this finding, direct interactions between the GABA and serotonin systems in the raphe nuclei and cortical regions have been reported [47,48]. A low basal level of GABA may result in the reduced facilitation of serotoninergic transmission in mood disorders [49] and suppression of serotonin activity is likely to disinhibit benzodiazepine behavior, manifesting anxiolytic actions [50]. Besides serotonin and GABA, other neurotransmitters have also been reported to be involved in the anxiolytic action of KRG. For example, Zhao et al [10] reported that KRG attenuated anxiety-like behaviors in rats during ethanol withdrawal through enhanced mesoamygdaloid dopamine system. Moreover, oral administration of ginsenoside Rb1 once daily for 14 consecutive days suppressed anxiety-like responses in a rat model of post-traumatic stress disorders, possibly through the modification of hypothalamic neuropeptide Y expression, tyrosine hydroxylase expression in the locus coeruleus, and hippocampal mRNA expression of BDNF [28]. To test the potential involvement of BDNF expression in the anxiolytic effect of KRG in immobilized mice, we also studied alterations in BDNF expression in the hippocampus. Our results showed that immobilization significantly reduced the mature BDNF levels in the hippocampus, but not in the cortex, and the reduced BDNF was effectively normalized by KRG (Fig. 6). It has been demonstrated that exposure to chronic stress abate mRNA and protein expression of BDNF in the limbic system, including the hippocampus [51,52]. Serum BDNF levels are also altered by several antidepressant drugs, such as FLX [52]. The serotonin and GABA systems and BDNF share common features of their abilities to regulate the development and plasticity of neural circuits involved in mood disorders, including depression and anxiety [53,54]. Our data revealed that p-CPA, flumazenil, or bicuculline dramatically inhibited the BDNF level elevated by KRG (Fig. 6), implying that enhanced BDNF expression in the hippocampus may be mediated by both serotonergic and GABAergic transmissions in the immobilized mice. Reciprocal interactions between BDNF and serotonin system have been demonstrated, in which BDNF promotes the development and function of serotonergic neurons expressing TrkB, the high-affinity receptor for mature BDNF [55]. The enhanced serotonergic transmission has been suggested to increase the production of cAMP, subsequently activating protein kinase A (PKA). The phosphorylation of cAMP response element-binding protein (CREB) by PKA may positively regulate the transcription of BDNF [56]. Further studies are in progress to examine whether the CREB-mediated signaling participates in the KRG-induced BDNF expression. The amygdala is another critical brain region involved in various types of stress-associated mental disorders like anxiety [57]. It would also be valuable to elucidate the effect of KRG on the BDNF expression in the amygdala of immobilized mice.
In summary, we demonstrated the anxiolytic effect of KRG in immobilized mice. Administration of KRG for 4 weeks markedly ameliorated immobilization-induced anxiety-like behaviors in mice, as assessed by the EPM and MBT. We also demonstrated that the serotonin and GABA systems and hippocampal BDNF expression were involved in the anxiolytic action. Upregulation of TPH and GABAA receptor expression by KRG may contribute, at least in part, to enhance the serotonin and GABA transmissions, respectively, in the brain. Interestingly, our data imply that the augmentation of BDNF level in the hippocampus may be associated with the enhanced serotonergic and GABAergic transmissions by KRG in the immobilized mice. Our findings would bring up new motives for elucidation of active compound(s) in KRG and molecular mechanisms underlying its anxiolytic effect. More efforts are required to clarify the connections between the serotonin and GABA systems and hippocampal BDNF expression. Collectively, KRG exhibited anxiolytic effect in the immobilized mice by multiple mechanisms of action, involving enhanced serotonin and GABA transmissions and BDNF expression. Considering the long-term safety profile of KRG, it may offer a promising alternative for anxiety treatment.
The authors declare that they have no conflicts of interest. | true | true | true |
|
PMC9597638 | Hongkang Zhu,Ruoyong Wang,Hanyi Hua,He Qian,Peng Du | Deciphering the potential role of Maca compounds prescription influencing gut microbiota in the management of exercise-induced fatigue by integrative genomic analysis | 12-10-2022 | anti-fatigue,gut microbiota,microbial function,oxidative stress,host metabolism | A growing number of nutraceuticals and cosmeceuticals have been utilized for millennia as anti-fatigue supplements in folk medicine. However, the anti-fatigue mechanism underlying is still far from being clearly explained. The aim of the study is to explore the underlying mechanism of the Maca compound preparation (MCP), a prescription for management of exercise-induced fatigue. In this study, mice weight-loaded swimming test was used to evaluate the anti-fatigue effect of MCP. MCP significantly improved the forelimb grip strength and Rota-rod test in behavioral tests via regulating energy metabolism. 16S rDNA sequencing results showed MCP can regulate the intestinal flora at the genus level by increasing several beneficial bacteria (i.e., Lactobacillus, Akkermansia and etc.), and decreasing the harmful bacteria (i.e., Candidatus_Planktophila and Candidatus_Arthromitus), where notable high relevance was observed between the fatigue-related biomarkers and fecal microbiota. The results of microbial function analysis suggested that MCP might improve exercise-induced fatigue by enhancing energy metabolism, carbohydrate and lipid metabolism and metabolism of terpenoids and polyketides and breakdown of amino acid metabolism. In addition, and H2O2-induced oxidative stress model on C2C12 cells was employed to further validate the regulation of MCP on energy metabolisms. MCP pre-treatment significantly reduced intracellular ROS accumulation, and increased glycogen content, ATP generation capacity and mitochondrial membrane potential of skeletal muscle cells, as well as conferred anti-cell necrosis ability. In conclusion, MCP plays a key role in regulating fatigue occurrence in exercising and gut microbiota balance, which may be of particular importance in the case of manual workers or sub-healthy populations. | Deciphering the potential role of Maca compounds prescription influencing gut microbiota in the management of exercise-induced fatigue by integrative genomic analysis
A growing number of nutraceuticals and cosmeceuticals have been utilized for millennia as anti-fatigue supplements in folk medicine. However, the anti-fatigue mechanism underlying is still far from being clearly explained. The aim of the study is to explore the underlying mechanism of the Maca compound preparation (MCP), a prescription for management of exercise-induced fatigue. In this study, mice weight-loaded swimming test was used to evaluate the anti-fatigue effect of MCP. MCP significantly improved the forelimb grip strength and Rota-rod test in behavioral tests via regulating energy metabolism. 16S rDNA sequencing results showed MCP can regulate the intestinal flora at the genus level by increasing several beneficial bacteria (i.e., Lactobacillus, Akkermansia and etc.), and decreasing the harmful bacteria (i.e., Candidatus_Planktophila and Candidatus_Arthromitus), where notable high relevance was observed between the fatigue-related biomarkers and fecal microbiota. The results of microbial function analysis suggested that MCP might improve exercise-induced fatigue by enhancing energy metabolism, carbohydrate and lipid metabolism and metabolism of terpenoids and polyketides and breakdown of amino acid metabolism. In addition, and H2O2-induced oxidative stress model on C2C12 cells was employed to further validate the regulation of MCP on energy metabolisms. MCP pre-treatment significantly reduced intracellular ROS accumulation, and increased glycogen content, ATP generation capacity and mitochondrial membrane potential of skeletal muscle cells, as well as conferred anti-cell necrosis ability. In conclusion, MCP plays a key role in regulating fatigue occurrence in exercising and gut microbiota balance, which may be of particular importance in the case of manual workers or sub-healthy populations.
Fatigue is a systemic process in the decline of human physiology and psychology, and most of the homeopathic supplements are holistic treatments. A growing number of edible natural plants have become used in dietary supplements for relieving fatigue-related symptoms (1, 2), where the compound prescriptions has become one of the most commonly used forms for exercise fatigue (3). Maca compound preparation (MCP) has been administered as an anti-fatigue agent in folk medicine in China for a long time according to The Medical Classic of the Yellow Emperor and Compendium of Materia Medica. In this formula, Maca (Lepidium meyenii Walp.), the monarch drug, is an edible medicine and a new resource food and its anti-fatigue effect has been investigated in our previous study (4). Macamides are the major constituents in Maca and their anti-fatigue activity is mainly credited to their antioxidant and anti-inflammatory properties (5). The eight edible and medicinal plants in MCP (Table 1) have been used to treat fatigue or weakness, which can also inhibit oxidative stress or inflammatory injury induced by exhausting physical exercise. With the in-depth study of intestinal flora, gut microbiota derived metabolites have been confirmed closely related to the progression of bidirectional communication pathways linking to host (6, 7). Therefore, the overall treatment targeting on both energy metabolism and intestinal bacteria may be a worthwhile therapeutic strategy to fatigue management. In our previous study, the active components of MCP and anti-fatigue potential have been predicted by network pharmacology exploration (8). However, the effect of MCP on anti-fatigue has not been fully explained and validated, and meanwhile, the association between fatigue-related effects and microbial function has not been established yet. In this study, the weight-loaded forced swimming test (WFST) was performed to verify its anti-fatigue effect, and the underlying mechanisms were systematically explored in exercise-induced fatigue mice. To further explore probable associations between its anti-fatigue effects and gut prebiotic capacity, we tried to measure some preliminary link between physical fatigue and gut microbiota in genus-level. The PICRUSt2 algorithm and MicroCyc databases (9, 10) were employed to investigate the metabolic functions of gut microbes and how MCP regulated gut microbiota and metabolic processes in the fatigue. Based on Kyoto Encyclopedia of Genes and Genomes pathway (KEEG) pathways analysis, metabolism pathways at level 1 are critically involved in host-microbial interactions, where energy metabolisms were significantly promoted by MCP. Finally, a typical H2O2-induced C2C12 cell injury model was applied to mimic skeletal muscle oxidative injury to investigate the protective effects of antioxidant pretreatment and validated the regulations of the intracellular energy metabolisms by MCP in vitro.
Lepidium meyenii Walp. was planted on the Tibet Plateau with an altitude over 3,000 meters. It was harvested in December, and the other 7 plants in MCP were harvested in Shandong Province, China, and were positively identified as the species by Guoliang Ding (Registered Traditional Chinese Medicine Practitioner, R.TCM.P., China). The main medicinal parts of plants in MCP (Table 1) were cut into thin sections (3–5 mm). After soaking for half an hour, MCP was extracted in boiling water (w/v, 1:8) twice for 1 h at atmospheric pressure. After filtration, the hot water extracts were combined and cooled to room temperature prior to centrifuge at 3,000 rpm for 10 min, further removing the suspended residue in the filtrate. MCP is of high nutritional value, containing 34.78 ± 2.43 mg/ml of total polysaccharides, 0.157 ± 0.018 mg/ml of flavonoids and 1845.27 ± 10.92 mg/ml of total amino acids according to the published paper which used the same protocol (8). The MCP extract was lyophilized at −70°C using a freeze dryer. All lyophilized extracts powder was sealed in sterile sampling bags and stored at −80°C for further experiments.
All experimental animal procedures were approved by the Ethics Committee of Experimental Animal Center of Jiangnan University (JN.No 20200710i0720915) and the whole project team was strongly committed to comply with the ethical policy. The experiments were performed when Institute of Cancer Research (ICR) mice (18–22 g, male) had adapted to the experimental environment for a week. Animals were randomly divided into 6 different groups (n = 10): control with vehicle treatment (Con); swimming exercise with vehicle treatment (Ex); swimming exercise with MCP (Ex + MCP) or caffeine (purity ≥ 99.8%, Pos, 10 mg/kg bw.) every day for 30 days. The vehicle group received the same volume of sterile water and the 1/10th of the recommended dosage for humans of MCP (12 g/day) received by the mice was based on the “mouse equivalent dose” (8, 18). Thus, MCP was orally administered at 1.0, 2.0, or 4.0 g/kg bw., respectively (the calculated process described in the Appendix), which was reconstituted in sterile water. The low, moderate, and high dose levels of MCP were labeled as MCP-L, MCP-M, MCP-H (Figure 1A).
The forelimb grip strength was assessed by a grip strength meter (YLS-13A Jinan Yiyan Technology Development Co., Ltd., Jinan, China) after oral gavage for 4 weeks. Each mouse was lifted by tail, so its forepaws could grip the wire of the strength meter, and then gently pulled back with tail parallel to table until the mouse lost its grip on the wire. Three tests were performed in succession on each mouse, and scores were averaged for statistical analysis.
Mice from each group were placed on an accelerating rota-rod cylinder (ZB-200, Chengdu Taimeng Science Technology Co., Ltd.). The rota-rod was accelerated from 5 to 15 rpm in 5 min. Each mouse remained on revolving rod for the training period. In the formal test, mice were placed on the rota-rod at speed of 15 rpm, until they were exhausted and dropped from the rod. The time of duration was recorded to evaluate motor coordination.
Weight-loaded forced swimming test (WFST) was carried out after orally administration on the last day (19, 20). The mice were placed in a tank (height: 65 cm; diameter: 40 cm) of room temperature water (25 ± 2°C) with a depth of 25 cm. The mice were loaded with a lead block (5% of body weight) attached to tails. After the WFST for 30 min, mice were sacrificed for the fatigue-related biochemical indexes measurement.
The blood was detected for blood sugar by Roche blood sugar meter (ACCU-CHEK Performa) and centrifugated at 3,500 g for 10 min at room temperature for measurement of blood lactic acid (BLA), blood urea nitrogen (BUN), lactate dehydrogenase activity (LDH) according to the instructions of the assay kit (Jiancheng Biotechnology Co., Nanjing, China). The left hind leg thigh muscle of mice was collected to measure the levels of reactive oxygen species (ROS), adenosine-triphosphate (ATP), nicotinamide adenine dinucleotide (reduce) [NAD(H)] and glycogen, which were detected (Fankew, Shanghai FANKEL Industrial Co., Ltd., Shanghai, China).
For histological examination, the muscle tissue was fixed in 4% (v/v) paraformaldehyde/PBS and embedded in paraffin, then stained the tissue with H&E. Finally, images were acquired from light microscopy (Olympus, Tokyo, Japan) (×200).
The 5 mice were selected from Ex and MCP-M groups at random, and DNA from feces was extracted by using a Genomic DNA Kit (Omega Bio-tek, Inc., Norcross, GA, USA). The 16S rRNA genes (V3-V4 regions) were amplified from the whole genome via the primer pair (341 F, 5′- CCTAYGGGRBGCASCAG-3′; 806R, 5′-GGACTACHVGGGTWTCTAAT-3′). All the amplicons were purified, quantified, and sequenced on an Illumina novaseq platform (San Diego, CA, USA). The barcode and connector sequence were removed. FLASH (v1.2.8) was used to stitch the double-ended sequences, and Vsearch (v2.3.4) was used to filter out the unqualified sequences (8). Finally, the sequence with 97% similarity was classified as an OTU. 16S rRNA gene reads were down-sampled to a read depth of 44,367 reads/sample and reads mapped to 16S OTUs (1,635 reads), to ensure sample compatibility regardless of sampling depth (21). Gene functions were predicted by PICRUSt.2 and analyzed by KEGG pathway enrichment analysis.
The C2C12 (mouse skeletal muscle) cells were purchased from (Hunan Fenghui Biotechnology Co., Ltd., Changsha, China). The cells were maintained in growth medium [Dulbecco’s modified Eagle medium (DMEM) supplemented, 10% (v/v) fetal bovine serum (FBS), 1% (v/v) Penicillin-Streptomycin for cell culture (Gibico)] in a 25 cm2 culture flask. The cells were cultivated at 37°C under a humidified atmosphere of 5% (v/v) CO2. To induce cell differentiation, 70% confluent cultures were switched to DMEM containing 2% horse serum (HS) and 10 μg/ml of insulin for 3 days with medium changes every other day.
When the C2C12 cells were grown to approximately 70–80% confluence in 96-well flat-bottomed plates (n = 5), they were replenished MCP (re-dissolved in basal medium, free of FBS), and incubated for a further 24 h. According to a previous study (4), cells were pre-treated with 0.48 mmol/L H2O2 for 6 h to inflict oxidative stress before harvest. The cell viability of reagents on C2C12 cells was, respectively, assessed using Cell Counting Kit-8 (CCK-8, Beyotime) assays. After incubating for 1h, the absorbance of each well was measured at 450 nm with a microplate reader (BioTek Instruments, Winooski, VT, USA).
The levels of ROS in C2C12 cells were determined using fluorescent probe (DCFH-DA, Beyotime, China). For flow cytometry analysis, C2C12 cells were seeded in 6-well plates at a density of 5 × 105 cells per well (treatment of cells was shown in 2.2.2). Both suspended and adherent cells were collected and washed with PBS twice (for washing out the trypsin). The cells were stained with DCFH-DA and incubated at dark ambient for 20 min. After re-suspending in PBS, they were analyzed by flow cytometry (FACSAriaII, Becton, Dickinson and Company, Franklin Lakes, NJ, USA) according to the manufacturer’s protocol. Analysis of flow cytometry data was performed with FlowJo software (v10.8.1).
To evaluate the effect of the compounds on energy metabolism during muscle growth, the anthrone reagent was employed in the estimation of glycogen content by use of glycogen assay kit (Solarbio Science & Technology Co. Ltd., Beijing, China). C2C12 cells were seeded in 6-wells plates, and 5 million cells would be harvested in 2 days. The glycogen content in cells was measured with an anthrone color reagent from an alkaline digest in concentrated sulfuric acid and calculated in the number of 104 cells. ATP levels in C2C12 cells were measured by an enhanced ATP assay kit (Beyotime Inc., Shanghai, China) as described in manufacturer’s instructions. Luminescence in the supernatant from each sample was measured in a Synergy Mx multifunctional Microplate Reader (Gene Company Ltd., Hongkong, China). Data were normalized to the control as 100%.
The Calcein-AM/PI Double Stain Kit (Dojindo Laboratories, Kumamoto, Japan) was used to assess the live/dead staining assay. In brief, the C2C12 cells were seeded onto glass coverslips in 96-well cell culture plates. Cells were treated as shown in 2.2.2. After co-incubation with agents (for 24 h) and H2O2 (for 6 h), cells were washed twice with 1 × Assay Buffer, then loaded with Calcein-AM and Propidium Iodide (PI) at 37°C for 20 min, washed twice in staining buffer, and fluorescence were measured by inverted fluorescence microscope (Carl Zeiss, Jena, Germany). The experiments were performed on C2C12 cells from 3 different visions at random.
Changes in ΔΨm of C2C12 cells were measured by using a mitochondrion-specific cationic dye JC-1 according to manufacturer’s instructions (Beyotime) (22). Cells were incubated with JC-1 working solution in the dark for 20 min, which was measured by high-resolution confocal laser microscope (LSM880, Carl Zeiss, Germany). Red-fluorescent emissions were formed by J-aggregates at high membrane potential (excitation 585 nm, emission 590 nm), whereas green-fluorescent monomers existed at low potential (excitation 514 nm, emission 529 nm). Mitochondrial depolarization manifests by a decrease in the ratio of the red and green fluorescence.
Swiss Target Prediction was used to predict all the targets of candidates, which were imported into Cytoscape 3.8.0 to construct a drug-target network. Molecular docking was carried out online by using SwissDock. Crystal structures of the four validated targets in PDB format were downloaded from AlphaFold Protein Structure Database and uploaded with candidate ligands in MOL2 format. The lowest Gibbs free energy (△G) and fullfitness of each interaction were calculated in silico.
GraphPad Prism 9.0 was used for analysis. Results were expressed as mean ± standard deviation. One-way ANOVA with Dennett’s multiple comparisons test was used for comparison between the three groups, *p < 0.05, **p < 0.01 were statistically significant among the groups. The correlation among these indicators was conducted by Pearson correlation analysis using R software version 4.1.0.
The imbalance of metabolic utilization in peripheral muscle may be both the cause and result of exercise-induced fatigue (23). To validate the effect of MCP on the physical fitness, the grip-strength test and Rota-rod test were carried out after administration of MCP for 4 weeks, as shown in Figure 1A. There is a significant improvement in the forelimb grip strength (Figure 1B), demonstrating MCP enhanced mice muscle strength force production. Meanwhile, the difference seemed to be implicitly stated in Rota-rod exercising (Figure 1C), since motor abilities on accelerating Rota-rod increased significantly in MCP mice. These results revealed improved locomotor capacity in mice treated with MCP. However, to determine whether MCP acted as an antagonist on fatigue/tiredness produced by exercise, we employed the WFST model to investigate the effects of MCP on muscle fatigue status in vivo.
To evaluate the energy metabolisms during fatigue, the contents of blood sugar, BLA, BUN, LDH in serum and NAD(H) in muscle were investigated in exercise mice. The occurrence of fatigue is accelerated by the decrease of blood sugar and accumulation of metabolites such as BLA, BUN and LDH, however, this situation is effectively reversed by the treatment of MCP as shown in Figures 2A–D. In addition to these metabolites, NAD(H) plays key role in the process of glycolysis process and cellular respiration, which is also an integral determining factor in fatigue (24). The levels of NAD(H) were decreased by approximately half in mice muscle after swimming, which were significantly increased by MCP-treatment compared with those of Ex group (Figures 2E,F). MCP increased energy materials in muscle, which could improve athletic capacity and enhance exercising endurance, in line with the observed results of exercise performance.
The above results of changes in plasma and muscle metabolites suggested abnormal energy metabolism of exercise mice during fatigue loading. To further explore the role of MCP in resistance to acute exercise-induced peripheral fatigue, we focused on the site of muscle in vivo. In WFST model, the ROS levels in the left hind thigh mice muscle were significantly down-regulated by MCP supplementation compared to the Ex group (p < 0.05) in Figure 3A, which is parallel to the outcomes of the in vitro tests. In addition, MCP concentration-dependently enhanced the muscle glycogen contents (Figure 3B) and ATP level (Figure 3C), promoting energy metabolism in muscle during swimming. To intuitively observe the protective effect of MCP, H&E staining were carried out on the hind leg thigh muscle (Figure 3D). Representative images showing H&E staining on mice muscle. Like the in vitro studies, the intense exercise led to muscle cells swelling, necrosis or degradation with abundant cracks. Pre-treatment of MCP at moderate or high dose exhibited pronounced cytoprotective effects, and protected muscle issues from disorganization and loss of tissue structure. Based on effects of the mice physical fitness, an obvious enhancement was observed on MCP at moderate dose compared with the low dose, which was larger than that between the moderate and high doses. Thus, the moderate dose (2.0 g/kg), which was equivalent to the recommended dose in humans, was selected for subsequent experiments.
According to the result of 16s RNA, the relative genus-level gut flora abundance was significantly different between Ex and Ex + MCP groups (Supplementary Table 1). MCP can improve resistance to fatigue by regulating the intestinal flora by increasing several primary beneficial bacteria (i.e., Lactobacillus, Akkermansia and etc.), and decreasing harmful bacteria (i.e., Candidatus_Planktophila and Candidatus_Arthromitus; Figure 4A). Notably, the relative genus-level gut flora abundance of Lactobacillus is increased over 2.5 folds (51.4 vs 19.8%, MCP vs Ex group), and that of Candidatus_Planktophila is decreased about 16 folds (2.2 vs 36.9%). Fatigue-related biochemical changes, as well as gut microbiota changes have been found their contributions to exercise mice performance (19, 25). In particular, fatigue generation is proved as one of the consequences of gastrointestinal imbalance, which may be associated with host metabolism and intestinal microecology (26). To further evaluate the correlations among oxidative stress and energy metabolism in serum and muscle, Pearson correlation is performed to analyze the correlations (Figure 4B). According to Pearson correlation analysis, high relevance was observed between fatigue-related biomarkers (BLD, BUN, LDH and ROS) and energy-supplying substances (blood sugar, MG and ATP), which were strongly correlated to fecal microbiota at the genus level as well (Figure 4C). Accumulating convincing evidence has shown that the gut flora plays key role in resistance to fatigue in clinic (25). According to matrix plots, the exercise performance is positively related to the relative abundance of Lactobacillus, blautia, Clostridia_UCG-014 ant etc. (r ≥ 0.5, p < 0.01), whereas negatively related to that of harmful bacteria (i.e., Candidatus_Planktophila and Candidatus_Arthromitus) (r ≤ −0.5, p < 0.01). The absolute values of Pearson correlation lower than 0.5 were hidden, which were accumulated and displayed in two-side. According to the cumulative value in Figure 4C, BLD is one of the most relative biomarkers indicating exercise-induced fatigue, indicating that BLD might play a key role in exercise physiological, performance, and energy utilization via microbe-host interactions (27). Meanwhile, Lactobacillus and Candidatus_Planktophila may be the most relative beneficial/harmful bacteria.
The host-microbial and microbe-microbial interactions are often governed by the complex exchange of metabolites, which are highly associated with the metabolic capacity of hosts in health and disease (28). In Figure 5A, networks of microbe-microbe interactions revealed the regulation of MCP among microbiomes (p < 0.01), where Lactobacillus suggested that it might occupy the dominant position due to the biggest circle size. To further investigate how MCP regulates gut microbiota and metabolic processes in the fatigue, the PICRUSt2 algorithm and MicroCyc databases were employed to analyze the metabolic functions of gut microbes. By comparing with the microbial genome sequences in the database, the key pathways (L1) involved in the breakdown of MCP by microbes were most related to metabolic pathways, and significantly distinguished from Ex in the environmental information processing, human diseases pathways and metabolism (Figure 5B). In addition, the gut microbial functions were analyzed and principal component (PCA) analysis revealed striking differences between the two groups based on KEGG orthology (L2; Figure 5C). Compared with the Ex group, MCP significantly regulated 19 pathways at levels 2 (Figure 5D) and the changes in the mice individual level were shown in Figure 5E. MCP could enhance the cellular processes of cell motility. Six human diseases were downregulated by MCP, involving in substance dependence, infectious disease, immune disease, drug resistance, cardiovascular disease and cancer. Additionally, treatment with MCP significantly reduced amino acid metabolism, and global and overview maps, while MCP significantly enhanced energy metabolism, carbohydrate, lipid metabolism, metabolism of terpenoids and polyketides compared with the Ex group. MCP might be critically involved in tissue development and tissue homeostasis in aging, circulatory system, digestive system, excretory system, immune system, and nervous system. Furthermore, MCP significantly downregulated the pathways of human cytomegalovirus infection, atrazine degradation, renin angiotensin system and etc. at level 3 (Supplementary Figure 1). Thus, the nutritional modulation of gut microbiota by MCP and its interplay between intestinal microbiota and host metabolism may provide a promising insight into fatigue process and a promising avenue for some metabolic dysfunctions (29).
Based on the above results, in vitro experiment was carried out to evaluate energy metabolism regulation of MCP in fatigue. According to the free radical theory, the accumulation of oxidative damage is a causal factor for muscle fatigue (30). Thus, H2O2 exposure was widely used for inducing C2C12 cell injuries to mimic oxidative damage in muscle fatigue (31, 32). Commonly, cell viability is a comprehensive index used to evaluate proliferation and cell death, indicating the potential of MCP to resist fatigue. In contrast, MCP (0.1–1.0 mg/ml) showed a marked potential for protecting skeletal muscle cells from oxidative stress injuries (Figure 6A). Oxidative stress and energy metabolism play indispensable roles in the in-tissue homeostasis and fatigue-related symptoms. The fluorescence intensity represented the amount of intracellular ROS production in C2C12 cells, which were detected by using DCFH-DA as fluorescent probe and analyzed by flow cytometry. H2O2 treatment increased ROS levels in cells, whereas MCP pre-incubation reduced H2O2-induced ROS generation (p < 0.01), suggesting that resistance to H2O2 injury was improved by MCP (Figure 6B). meanwhile, MCP concentration-dependently enhanced the intracellular glycogen and ATP contents, promoting energy metabolism in C2C12 cells. Under oxidative stress damage, the glycogen contents in Mod decreased significantly (**p < 0.01) by nearly 30%. The MCP groups showed higher contents of C2C12 glycogen and ATP levels than those of Mod groups, indicating its anti-fatigue potential on H2O2 mediated-oxidative stress on cells. Moreover, it was worth noting that MCP concentration-dependently enhanced the glycogen content (R2 = 0.9983), and ATP generation ability (R2 = 0.9917) in vitro (Figures 6C,D). MCP significantly conferred anti-cell necrosis ability and attenuated exercise-induced damage, as evidenced by an increase of Live/Dead assay and skeletal muscle morphology (Figure 6E). Compared with Con, the H2O2 shrunk the cells, deformed their shapes, and generated large quantities of cell debris, while MCP significantly alleviated H2O2-induced oxidative damage both in viability or cell morphology. As shown in Figure 6F, H2O2 injury impaired ΔΨm, while MCP treatment improved the stability of ΔΨm and morphology, indicating protective effects on mitochondrial functions. Thus, the in vitro experiments further demonstrated a substantial effect of MCP on mitochondrial function and energy metabolism.
Mitochondrial respiratory metabolism is the major source of cellular glycogen and ATP, and is associated with the proper maintenance of cellular metabolism as a whole (33). To further explore the key metabolic pathways of MCP, a joint pathway analysis was performed based on network pharmacology and metabolic pathway analyses on MetaboAnalyst platform. Eight anti-fatigue intestine-specific expressed targets (ABCG2, PDE9A, SLC6A4, CHRNA7, HNF4A and MAOA) of MCP and six metabolites (BLA, blood sugar, glycogen, ATP, NAD and NADH) were plugged into MetScape to build compound-reaction-enzyme-gene networks by matching with the potential targets obtained from MetScape (Figure 7A). The intestine-specific expression of phosphodiesterase 9A (PDE9A) was identified as the key target. Therefore, we believe that PDE9A may have a crucial effect on the efficacy of MCP for fatigue. Several bioactive substances might play critical roles in MCP. Based on previous network pharmacology analysis (8), N-benzyl-octanamide (from Maca), vitamin E (from Amomum villosum Lour), cnidilin and phyllanthin (from Angelica sinensis), obacunone (from Citrus reticulata Blanco), are the active ingredients targeted on PDE9A. Figure 7B showed the visualization of the most energetically favorable binding of the six ligands into the key protein of PDE9A. As the most favorable interaction, cnidilin and obacunone showed the lowest Gibbs free energy (△G = −6.36 and −6.11 kcal/mol) and for PDE9A, suggesting that suggesting that dietary influences by MCP are most probably mediated in part by the gut microbes.
Fatigue were the most commonly reported systemic reactions, suggesting a link existed among fatigue-related parameters from central to peripheral (19). Intestinal dysbiosis can dysregulate inflammation of intestinal surrounding tissues as well as cognition and mood (34). In the formula, Maca could improve male sexual behavior so that it has been known as “South American ginseng” or “Plant Viagra” worldwide (5). Its extract could increase glucose uptake by inhibiting mitochondrial function (35). Natural herbal medicine with compatibility and multiple targets, is widely regarded as a new way for drugs discovery for anti-fatigue. It points out directions for future studies on the internationalization of Traditional Chinese Medicine (TCM) (36). Under complementarity principle, the prescription was expected to exert anti-fatigue activities on multiple targets with the compatibility detoxication-based TCM theory. Meanwhile, the combined extracts produced better synergistic effects than a single drug (37). Compared with conventional single-target drugs, TCM prescriptions usually contain several medicinal herbs, along with synergistic effects (multiple compounds – multiple targets – one pathway) (38, 39). Increasing numbers of evidence suggested that gut microbiota-targeted therapy is a promising strategy to treat the chronic or metabolic disorders, which may behave as one of the mechanism for altering the pathogenesis TCM herbals (40). Currently, accumulated results have showed that Lactobacillus could significantly elevated the exercise performance in a dose-dependent manner and improved the fatigue-associated features correlated with better physiological adaptation (41). A recent study has revealed the impacts of gut microbiome on the host’s health, where Lactobacillus primarily modulated the overall microbial community structure (42). On the other hand, Candidatus Planktophila is an actinobacterium, which might represent pathogenic bacteria from freshwater bacterioplankton (43). Candidatus_Arthromitus was proved associated with depression (44). PDE9A is an intracellular cyclic guanosine monophosphate (cGMP) hydrolase, which has been exploited as one of the most promising therapeutics for treatment of metabolic diseases, such as diabetes and Alzheimer’s (45), as well as fatigue syndromes (46). In this study, the link between fatigue-related parameters and gut microbiota have been established by treatment of MCP, however, the underlying mechanisms should be further explored – the gut microbiota and the host interplay. Thus, further studies on host-microbe-drug-nutrient may probably need to answer the fundamental question of how MCP maintains intestinal microecological balance and delays generation of host fatigue (47). In conclusion, it was demonstrated a medicinal and edible decoction – MCP, as a promising candidate to manage exercise-induced fatigue, which may be of particular importance in the case of manual workers or sub-healthy populations.
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and number(s) can be found below: NCBI [BioProject: PRJNA888212 for 16S rRNA sequencing].
The animal study was reviewed and approved by the Ethics Committee of Experimental Animal Center of Jiangnan University (JN.No 20200710i0720915).
HZ: conceptualization and writing – original draft preparation. RW and HH: validation and formal analysis. HQ and PD: conceptualization. All authors contributed to the article and approved the submitted version. | true | true | true |
|
PMC9597811 | Wenxiu Ru,Kunpeng Liu,Jiameng Yang,Jianyong Liu,Xinglei Qi,Bizhi Huang,Hong Chen | miR-183/96/182 Cluster Regulates the Development of Bovine Myoblasts through Targeting FoxO1 | 17-10-2022 | bovine myoblasts,miR-183/96/182 cluster,FoxO1,proliferation,differentiation | Simple Summary In this work, we identified that the miR-183/96/182 cluster was highly expressed in bovine embryonic muscle; meanwhile, it widely existed in other organizations. Functional assays indicated that the miR-183/96/182 cluster targets the FoxO1 gene to regulate the proliferation and differentiation of bovine myoblasts. Abstract Muscle development is an important factor affecting meat yield and quality and is coordinated by a variety of the myogenic genes and signaling pathways. Recent studies reported that miRNA, a class of highly conserved small noncoding RNA, is actively involved in regulating muscle development, but many miRNAs still need to be further explored. Here, we identified that the miR-183/96/182 cluster exhibited higher expression in bovine embryonic muscle; meanwhile, it widely existed in other organizations. Functionally, the results of the RT-qPCR, EdU, CCK8 and immunofluorescence assays demonstrated that the miR-183/96/182 cluster promoted proliferation and differentiation of bovine myoblast. Next, we found that the miR-183/96/182 cluster targeted FoxO1 and restrained its expression. Meanwhile, the expression of FoxO1 had a negative correlation with the expression of the miR-183/96/182 cluster during myoblast differentiation. In a word, our findings indicated that the miR-183/96/182 cluster serves as a positive regulator in the proliferation and differentiation of bovine myoblasts through suppressing the expression of FoxO1. | miR-183/96/182 Cluster Regulates the Development of Bovine Myoblasts through Targeting FoxO1
In this work, we identified that the miR-183/96/182 cluster was highly expressed in bovine embryonic muscle; meanwhile, it widely existed in other organizations. Functional assays indicated that the miR-183/96/182 cluster targets the FoxO1 gene to regulate the proliferation and differentiation of bovine myoblasts.
Muscle development is an important factor affecting meat yield and quality and is coordinated by a variety of the myogenic genes and signaling pathways. Recent studies reported that miRNA, a class of highly conserved small noncoding RNA, is actively involved in regulating muscle development, but many miRNAs still need to be further explored. Here, we identified that the miR-183/96/182 cluster exhibited higher expression in bovine embryonic muscle; meanwhile, it widely existed in other organizations. Functionally, the results of the RT-qPCR, EdU, CCK8 and immunofluorescence assays demonstrated that the miR-183/96/182 cluster promoted proliferation and differentiation of bovine myoblast. Next, we found that the miR-183/96/182 cluster targeted FoxO1 and restrained its expression. Meanwhile, the expression of FoxO1 had a negative correlation with the expression of the miR-183/96/182 cluster during myoblast differentiation. In a word, our findings indicated that the miR-183/96/182 cluster serves as a positive regulator in the proliferation and differentiation of bovine myoblasts through suppressing the expression of FoxO1.
Skeletal muscle is the most abundant tissue accounting for most of the body weight and participates in movement and metabolism [1]. For livestock, skeletal muscle development is an important factor affecting meat yield and quality. Skeletal muscle originates from mesenchymal stem cells (MSC) in the embryonic mesoderm, which could change to myogenic progenitor cells. The myogenic progenitor cells differentiate into mononuclear myoblasts; then, the myoblasts further experience proliferation, differentiation and fusion into multinuclear myotubes, which ultimately form muscle fibers [2]. Once the mature muscle has formed, myogenic progenitor cells will enter quiescence and exist as muscle satellite cells, which could participates in the repair of muscle fibers [3]. It is now generally accepted that this complex and long-term process is precisely coordinated by the myogenic regulatory factors (MRFs), containing Myogenic differentiation 1 (MyoD1), Myogenic regulatory factor 4 (Mrf4), Myogenin (MyoG) and Myogenic factor 5 (Myf5) [4]. Tuning these genes’ expression in muscle development is realized through a transcriptional and post-transcriptional network. Increasing evidence has suggested that noncoding RNAs (ncRNAs) involving post-transcriptional regulation is a vital factor in muscle development, including long noncoding RNA [5], circular RNA and small RNA [6,7]. Although the genetic and molecular pathways of regulating muscle development have been well-established in the past decades, numerous unknown regulatory molecules and mechanisms involved in this process remain unidentified. MicroRNAs (miRNA), a class of ~22 nucleotide small noncoding RNAs, are highly conserved and do not possess potential coding. The seed region of miRNA located in 2~8 nucleotides at the 5′ end normally conjugates with the 3′-untranslated region (3′ UTR) of the target mRNA. By this means, miRNAs are capable of decaying or impeding protein translations of the target mRNAs [8]. Numerous studies have underlined the key roles of miRNAs in regulating skeletal muscle development through their inhibiting effects on several myogenic regulatory factors and important signaling pathways. The MyomiR family, a class of miRNA specifically expressed in muscles, includes miR-133, miR-1, miR-208a/b, miR-206, miR-486 and miR-499 [9]. For instance, miR-206, which is highly and exclusively expressed in muscles, was positively regulated by MyoD and targeted Pax7 to promote terminal differentiation [10], while some non-muscle-specific miRNAs also participated in regulating muscle development, such as miR-24, miR-27a, miR-125b, miR-29, miR-486, miR-221/222 and miR-214 [11]. Moreover, the indispensable functions of miRNAs in muscle development have verified that the deficiency of Dicer in skeletal muscle reduced the production of muscle miRNAs and led to embryonic death during embryogenesis [12]. The conserved miR-183/96/182 cluster is one of the most studied miRNA clusters, which is situated at a 5-kb genomic region in humans and mice and possesses similar seed sequences [13]. Many upstream regulators, including Wnt/beta-Catenin, the p21-ZEB1 complex, GSK3β, MyoD and so on, were certified to regulate the expression of the miR-183/96/182 cluster [14,15,16,17]. Increasing studies have shown that the miR-183/96/182 cluster plays an important role in tumorigenesis, cancer progression, tumor invasion and metastasis. Notably, an inconsistency has arisen with respect to the functions of the miR-183/96/182 cluster in various tumor cells [18], which prompted that its versatility in different biological processes. Nevertheless, there are research gaps in the regulatory role of the miR-183/96/182 cluster in muscle development, especially in livestock animals. Therefore, we attempted to explore the function and mechanism of the miR-183/96/182 cluster in bovine skeletal muscle development. In this study, we found that the miR-183/96/182 cluster accelerated the proliferation and differentiation of bovine myoblasts by targeting the FoxO1 gene, which enriched the network of miRNAs regulating bovine muscle development.
Muscle samples in Qinchuan cattle were collected at the adult stage (24 months) and embryonic stage (90 days) from a local livestock farm in Xi’an (Shannxi, China). Other tissue samples included a liver, kidney, heart, lung and spleen that were obtained at the embryonic stage (90 days), and each sample was collected from three cattle about the same age. After surgical removal, all samples were placed in liquid nitrogen to snap-freeze and then kept at −80 °C until RNA isolation. Our study protocols were approved by the Animal the Ethics Committee of Northwest A&F University.
As previously described, the primary myoblasts of bovines were obtained from the longissimus muscle at the fetal stage (90 days) [19]. Briefly, the stripped muscle tissues were cut into pieces, then digested with collagenase I in 37 °C for 2 h, whereafter the digested muscles were filtered using a 200-mesh filter, and the filtrate was washed with PBS and centrifuged at 1500 rpm three times. Finally, the collected cells were cultured in DMEM (BI, ISR) supplemented with 20% fetal bovine serum (BI, ISR) and 2% penicillin–streptomycin (Biosharp, Anhui, China). When 90% confluence was reached, the medium was changed to DMEM with 2% horse serum and 2% penicillin–streptomycin to induce cell differentiation. The HEK-293T cells were cultured in DMEM with 10% FBS and 1% penicillin–streptomycin. All cells were cultured in a humidified incubator at 37 °C with 5% CO2.
The fragment of bovine FoxO1 3′UTR containing the miR-183/96/182 cluster wild binding site or mutant sites was cloned into the psiCHECK2 vector. These vectors were verified by sequencing. The mimics and inhibitors of the miR-183/96/182 cluster were synthesized by General Biol (Chuzhou, China) to the overexpression or knockdown of the miR-183/96/182 cluster, respectively. All vectors and mimics or inhibitors were transfected into myoblasts using Troubfect (Thermo Fisher Scientific, Waltham, MA, USA).
The total RNA from tissues and cells was extracted using the Trizol reagent (AG, Beijing, China) and was reverse transcribed using the RT reagent kit (AG, Beijing, China). For mRNA, random and oligo (dT) primers were used to synthesize cDNA, and for miRNA, random and stem-loop primers were used to synthesize cDNA. The RT-qPCR assay was performed using the SYBR Green Kit (Vazyme, Nanjing, China) on the CFX96 System (Bio-Rad, Hercules, CA, USA). We applied β-actin and U6 as an internal control for the mRNA and miRNA. The quantitation data was analyzed by the 2−ΔΔCt method, and each sample was replicated three times. All primers are listed in Table S1.
The myoblasts were cultured in 96-well plates and then transfected after the cell density reached 70–80%. For the CCK-8 assay, each well was treated with 10 μL of CCK-8 (UE, Suzhou, China) and incubated for 2 h at 37 °C in the dark. The absorbance of each sample was detected using a microplate reader at 450 nm, and each sample was replicated eight times. The EdU Cell Proliferation Kit (Beyotime, Shanghai, China) was used to measure the capacity of cell proliferation, and Hoechst 33342 was used to stain the nuclei. Each sample was replicated three times. Finally, the images were observed by fluorescence microscopy (AMG EVOS, SEA, USA).
After transfection, bovine myoblast differentiation was induced for 4 d. Then, the cells were washed with PBS and fixed with 4% paraformaldehyde for 30 min. After washing, the cells were permeabilized with 0.5% Triton X-100 for 15 min and blocked with 5% BSA for 30 min, following incubation at 4 °C overnight with antibody-MyHC diluted 1:250 (GeneTex, Irvine, CA, USA). Next, we used the homologous fluorescent secondary antibody (Immunoway, Plano, TX, USA) diluted 1:500 to incubate cells for 2 h at room temperature. Hoechst 33342 was used to stain the nuclei. Finally, the images were observed under a fluorescent microscope.
The psiCHECK2-FoxO1-WT or psiCHECK2-FoxO1-Mut and miR-183/96/182 cluster mimics were co-transfected into HEK293T cells using Troubfect (Thermo Fisher Scientific, Waltham, MA, USA). After transfection for 24 h, the cells were lysed, and next, we used the Dual-Luciferase Reporter Assay Kit (Promega, MDN, USA) to detect luciferase activities according to the manufacturer’s instructions. Finally, the ratios of renilla and firefly activity were calculated, and each sample was replicated eight times.
We used GraphPad Prism 8.0 (GraphPad, San Diego, CA, USA) and SPSS 22.0 (SPSS, Chicago, IL, USA) to analyze the data. Significance analyses between two groups were performed by an independent sample t-test, and for three or more groups, one-way analysis of variance (ANOVA) was used to compare any discrepancies. All data were presented as the mean ± SEM. A statistical significance was indicated as * p < 0.05, ** p < 0.01, *** p < 0.001 or **** p < 0.0001.
There have been numerous studies that have proven the important roles of the miR-183/96/182 cluster in cell proliferation, apoptosis and differentiation in various cancer cells [18]. In this study, we aimed to explore whether the miR-183/96/182 cluster also has regulatory functions in bovine myoblasts. Firstly, we downloaded the mature sequences of miR-183, miR-96 and miR-182 to examine their conservatism in different species. The results demonstrated that the miR-183/96/182 cluster is highly conserved among species (Figure 1A). Through quantifying the expression features of the miR-183/96/182 cluster, we found that it exhibits varying expression patterns in various tissues in bovines (Figure 1B). In addition, the miR-183/96/182 cluster had a significantly higher presence in embryonic muscle (Figure 1C), which suggested that it may be a positive regulator in embryonic muscle development.
The miR-183/96/182 cluster has been shown to promote most cancer cell proliferation [18]. To verify whether the miR-183/96/182 cluster influences bovine myoblast proliferation, we transfected bovine myoblasts with miR-183/96/182 mimics to increase its expression. The expression level of miR-183/96/182 was significantly higher than the control group (Figure 2A). A RT-qPCR assay was used to detect the expression of marker genes of proliferation. Our results demonstrated that the overexpression of the miR-183/96/182 cluster remarkably enhanced the expression of PCNA, CDK2 and cyclin D; meanwhile, the level of P21 was significantly reduced (Figure 2B). The EdU proliferation assays revealed that the overexpression of miR-183/96/182 significantly increased the number of EdU-positive cells, and silencing miR-183/96/182 reduced the positive cells (Figure 2C,D). In addition, the results of the CCK8 assays showed that overexpressing miR-183/96/182 could increase the vitality of myoblasts, and silencing miR-183/96/182 could inhibit the vitality of the myoblasts (Figure 2E). Thus, our results showed that the miR-183/96/182 cluster can facilitate the proliferation of bovine myoblasts.
Next, we verified the potential functions of the miR-183/96/182 cluster in myoblast differentiation. After the transfection of miR-183/96/182 mimics for 24 h, we used 2% horse serum medium to induce myoblast differentiation until day 3. Subsequently, we detected the expression level of the differentiation marker genes by RT-qPCR. As shown in Figure 3A, there is a significant rise in the expression level of the differentiation marker genes, including MyoG, MyoD and Myf5. However, silencing the miR-183/96/182 cluster reduced the expression of differentiation marker genes at the mRNA level. Additionally, the immunofluorescence assay also demonstrated that overexpressing the miR-183/96/182 cluster could prompt the myotube to become bigger (Figure 3B). Inversely, the myotube became lesser after silencing the miR-183/96/182 cluster (Figure 3B). Consequently, all the results suggested that the miR-183/96/182 cluster promoted bovine myoblast differentiation.
FoxO1, a well-known direct target of the miR-183/96/182 cluster, has been confirmed in various types of cancers, such as prostate [20], liver [16], breast [21], lymphoma [22] and so on. In addition, FoxO1 has also been proven to participate in muscle differentiation and glucose and lipid metabolism in skeletal muscle [23,24,25]. However, whether the miR-183/96/182 cluster regulates FoxO1 in bovine muscle is still unclear. As expected, we found that FoxO1 was potentially targeted by the miR-183/96/182 cluster, and the distribution of potential binding sites is demonstrated in Figure 4A. To investigating the binding of the miR-183/96/182 cluster and FoxO1, we constructed a dual-luciferase reporter system of FoxO1 3′UTR (wild type) (Figure 4B). The results indicated that the luciferase activity of pCK-FoxO1-WT was notably suppressed in HEK293T cells after co-transfection with the miR-183/96/182 cluster mimics (Figure 4C). Analogously, we further used the vector of pCK-FoxO1-MUT to verify this interaction (Figure 4A,D). There was no longer a response in the pCK-FoxO1-MUT system when the miR-183/96/182 cluster mimics were transfected (Figure 4D). The RT-qPCR assay indicated that the expression of FoxO1 was markedly decreased after overexpressing the miR-183/96/182 cluster in bovine myoblasts (Figure 4E). Additionally, we found that the expression of FoxO1 had a negative correlation with the expression of the miR-183/96/182 cluster during myoblast differentiation (Figure 4F). Taken together, these results demonstrated that the miR-183/96/182 cluster regulates the proliferation and differentiation of bovine myoblasts by targeting FoxO1.
Muscle development is precisely coordinated by members of the myocyte enhancer factor 2 (MEF2) family and myogenic regulatory factors (MRFs) [26,27]. In recent years, miRNA have been certified to mediate the post-transcriptional regulation of gene expression by RNA interference [13]. To date, numerous studies have established a powerful role of miRNAs in cell differentiation, growth, apoptosis and development, as we all know that myoblast proliferation and differentiation are the key factors affecting muscle development. Therefore, the potential effects of miRNAs in myoblast proliferation and differentiation cannot be ignored. As we all know, the conserved miR-183/96/182 cluster is one of the most studied miRNA clusters, which possesses similar seed sequences to target identical genes; meanwhile, it plays a crucial role by coordinating the key genes in various cellular processes [13]. Consistent with our analysis, the mature fragment of the miR-183/96/182 cluster is highly conserved in different species (human, mouse, bovine, rat, pig and chicken). However, it is still unclear whether the miR-183/96/182 cluster has a regulatory function in bovine myoblasts. In this study, we indicated that the miR-183/96/182 cluster serves as a positive regulator in the proliferation and differentiation of bovine myoblasts through suppressing the expression of FoxO1. Previous studies have demonstrated that the miR-183/96/182 cluster possessed an accelerated role in cell proliferation in most types of cancer. A classic example is that miR-182 and miR-183 facilitate cell proliferation and tumor invasion by inhibiting PDCD4 in various cancer cells, which is a typical tumor suppressor gene [28,29]. Interestingly, miR-96 was found to weaken pancreatic cancer cell proliferation, and miR-183 suppressed gastric cancer proliferation in the past few years [30,31]. These discrepant results may be caused by different types of cancer cells or competition between target genes or the involvement of different signaling pathways. Therefore, it is necessary for us to continue to explore the regulatory functions of the miR-183/96/182 cluster in various cellular environments and different life processes. Recent studies have reported the modulating capability of miR-96 and miR-183 in muscle oxidative, and the results showed that miR-96 and miR-183 can inhibit glucose utilization and fat catabolism [32]. Here, we further exhibited that the miR-183/96/182 cluster plays a vital role in promoting skeletal muscle proliferation and differentiation. The miR-183/96/182 cluster exhibited widely different expression patterns across bovine tissues, which is also consistent with its extensive functions in various tissues and cell development. Specifically, we observed a higher level of the miR-183/96/182 cluster in embryonic muscle tissue than adult muscle. The embryonic period is a critical stage for muscle growth and development; in addition, the miR-183/96/182 cluster can accelerate the proliferation and differentiation of bovine myoblasts, suggesting that miR-183/96/182 plays a crucial part in myogenesis. Notably, several studies in mice have confirmed that miR-96-5p and miR-183-5p, via suppressing FHL1, impede the differentiation and fusion of myoblasts [33,34]. Contrary to their results, we found that the miR-183/96/182 cluster promoted the proliferation and differentiation of bovine myoblasts through using RT-qPCR, EdU, CCK8 and immunofluorescence assays. This positive regulatory effect of the miR-183/96/182 cluster for bovine myoblasts is essential for skeletal muscle development. To be noted, the differences between species are likely mainly responsible for the function differences. Additionally, there may be competition between target genes in a specific context. Mechanism studies have shown that FoxO1 acts as a target gene of the miR-183/96/182 cluster to mediate bovine myoblast development. It is known that FoxO1, a member of the “O” subclass of the (FOX) family, has a considerable effect on various cellular physiological process. FoxO1 is found in most muscle types and plays a vital regulation in myoblast proliferation and differentiation, muscle growth and metabolism [35]. The function of FoxO1 in muscle differentiation has been widely reported. Some researchers have demonstrated that FoxO1 facilitates myotube fusion in mouse primary myoblasts [36]. However, other studies have revealed that FoxO1 is an inhibitor for muscle differentiation. At an early myogenesis stage, FoxO1 has been reported to impinge on the nutrient-sensing mTOR pathway, the Notch pathway and myostatin to repress myoblast differentiation [37,38,39]. The study in vivo showed that FoxO1 transgenic mice exhibited a remarkable reduction in muscle mass and the decreased expression of type I fiber genes and impaired skeletal muscle production [25]. Generally, most studies support FoxO1 as an inhibitor in myogenesis. Furthermore, increasing studies have verified that FoxO1 is a direct target gene to the miR-183-96-182 cluster. In bovine ovaries, the miR-183-96-182 cluster promotes granulosa cell proliferation, cycle progression and restrains apoptosis through targeting FoxO1 [40,41]. In the present study, we also confirmed that FoxO1 was the target gene of the miR-183-96-182 cluster using the dual-luciferase reporter assay, and overexpressing miR-183/96/182 cluster led to decreasing levels of FoxO1 mRNA in bovine myoblasts. Importantly, we observed that the expression of FoxO1 was negatively correlated with the expression of the miR-183/96/182 cluster during myoblast differentiation. Thus, our results established a pattern that the miR-183/96/182 cluster regulates the proliferation and differentiation of bovine myoblasts by targeting FoxO1. It is worth noting that we just uncovered a potential mechanism that the miR-183/96/182 cluster regulates muscle development, but there are other target genes of the miR-183/96/182 cluster that unceasingly need to be elucidated.
Overall, we characterized the miR-183/96/182 cluster in bovine muscle and identified the miR-183/96/182 cluster as a positive regulator in the differentiation and proliferation of bovine myoblasts. Mechanistically, we found that the miR-183/96/182 cluster targets the FoxO1 gene to regulate the proliferation and differentiation of bovine myoblasts. Our findings not only confirmed the universality of the regulatory functions of the miR-183/96/182 cluster in various biochemical processes but also provided a theoretical basis to clarify skeletal muscle development in bovines from a layer of noncoding RNAs. | true | true | true |
|
PMC9597991 | Lin Zhu,Xiaoyu Wang,Fuyi Li,Jiangning Song | PreAcrs: a machine learning framework for identifying anti-CRISPR proteins | 25-10-2022 | Anti-CRISPR protein,Machine learning,Feature engineering,Sequence analysis | Background Anti-CRISPR proteins are potent modulators that inhibit the CRISPR-Cas immunity system and have huge potential in gene editing and gene therapy as a genome-editing tool. Extensive studies have shown that anti-CRISPR proteins are essential for modifying endogenous genes, promoting the RNA-guided binding and cleavage of DNA or RNA substrates. In recent years, identifying and characterizing anti-CRISPR proteins has become a hot and significant research topic in bioinformatics. However, as most anti-CRISPR proteins fall short in sharing similarities to those currently known, traditional screening methods are time-consuming and inefficient. Machine learning methods could fill this gap with powerful predictive capability and provide a new perspective for anti-CRISPR protein identification. Results Here, we present a novel machine learning ensemble predictor, called PreAcrs, to identify anti-CRISPR proteins from protein sequences directly. Three features and eight different machine learning algorithms were used to train PreAcrs. PreAcrs outperformed other existing methods and significantly improved the prediction accuracy for identifying anti-CRISPR proteins. Conclusions In summary, the PreAcrs predictor achieved a competitive performance for predicting new anti-CRISPR proteins in terms of accuracy and robustness. We anticipate PreAcrs will be a valuable tool for researchers to speed up the research process. The source code is available at: https://github.com/Lyn-666/anti_CRISPR.git. Supplementary Information The online version contains supplementary material available at 10.1186/s12859-022-04986-3. | PreAcrs: a machine learning framework for identifying anti-CRISPR proteins
Anti-CRISPR proteins are potent modulators that inhibit the CRISPR-Cas immunity system and have huge potential in gene editing and gene therapy as a genome-editing tool. Extensive studies have shown that anti-CRISPR proteins are essential for modifying endogenous genes, promoting the RNA-guided binding and cleavage of DNA or RNA substrates. In recent years, identifying and characterizing anti-CRISPR proteins has become a hot and significant research topic in bioinformatics. However, as most anti-CRISPR proteins fall short in sharing similarities to those currently known, traditional screening methods are time-consuming and inefficient. Machine learning methods could fill this gap with powerful predictive capability and provide a new perspective for anti-CRISPR protein identification.
Here, we present a novel machine learning ensemble predictor, called PreAcrs, to identify anti-CRISPR proteins from protein sequences directly. Three features and eight different machine learning algorithms were used to train PreAcrs. PreAcrs outperformed other existing methods and significantly improved the prediction accuracy for identifying anti-CRISPR proteins.
In summary, the PreAcrs predictor achieved a competitive performance for predicting new anti-CRISPR proteins in terms of accuracy and robustness. We anticipate PreAcrs will be a valuable tool for researchers to speed up the research process. The source code is available at: https://github.com/Lyn-666/anti_CRISPR.git.
The online version contains supplementary material available at 10.1186/s12859-022-04986-3.
CRISPR-Cas adaptive immune system is one of the most widespread immunity strategies in prokaryotes against invading bacteriophages and plasmids [1, 2]. To counteract and overcome different CRISPR-Cas immunity systems, bacteriophages have evolved anti-CRISPR proteins (Acrs) that were first discovered in Pseudomonas aeruginosa phages in 2013 [3]. Subsequently, a proliferation of Acrs has proved to inactivate multiple CRISPR subtypes [3–7]. Several methods have been proposed to identify Acrs, including “Guilt-by-association” studies [6, 8], self-targeting CRISPR arrays [6, 7], and metagenome DNA screening [9, 10], etc. These methods assumed the new Acrs are similar to the previous Acrs. However, most Acrs fall short in sharing similarities currently acknowledged. Therefore, the traditional screening methods based on homology search are unreliable and require a lot of prior knowledge of Acrs to identify new Acrs. For instance, the “Guilt-by-association” method involves searching for homologs of helix-turn-helix (HTH)-containing proteins that are typically encoded downstream of Acrs [11]. The performance of “Guilt-by-association” is unstable when known Acrs proteins might share low similarity with queried protein. Therefore, a computational approach with less requirement for prior knowledge of known Acrs will provide a new perspective on the identification of Acrs. Machine learning algorithms with appropriate features could reveal the potential mechanism of Acrs and identify the Acrs without prior knowledge. Recently, some machine learning methods have been presented for predicting Acrs. There are several web servers about Acrs, such as: Anti-CRISPRdb [12], AcrHub [13], AcrDB [14], CRISPRminer2 [15], AcRanker [14, 16], AcrFinder [17], AcrCatalog [18] and PaCRISPR [19]. Anti-CRISPRdb, AcrDB, and AcrCatalog are online Acr datasets, while AcrHub, CRISPRminer2, AcRanker, AcrFinder and PaCRISPR are prediction web servers. Eitzinger et al. developed AcRanker, using the XGBoost ranking model to predict candidate Acrs only based on protein sequence information [16]. Wang et al. proposed PaCRISPR, an ensemble learning-based predictor, to identify Acrs from protein datasets derived from genome and metagenome sequencing projects [19]. Gussow et al. proposed a machine learning approach, using a random forest model with extremely randomized trees to expand the repertoire of Acrs families [20]. These machine learning methods have made a great contribution to discovering Acrs. However, the most appropriate features or feature combinations for Acrs prediction have not been systematically assessed. For instance, The PaCRISPR method identified the Acrs using only evolutionary features, and the AcRanker used only amino acid composition features to identify Acrs. Gussow et al. predict Acrs based on the sequence alignment and a heuristic secondary screen of few known Acrs. Thus, since previous work did not fully assess the feature combinations and relied on prior knowledge, we proposed a novel, effective and robust machine learning framework to help identify Acrs. This study presented an ensemble machine learning method, called PreAcrs, to efficiently and accurately predict Acrs based on protein sequences. Specifically, we used three features and eight different machine learning methods to train our model. 412 experimentally validated Acrs and 412 non-Acrs were introduced in the training dataset, and 176 were experimentally determined Acrs and 176 non-Acrs in the independent dataset. We found that the PreAcrs method outperformed other existing predictors with an AUC of 0.972 in the independent dataset.
To find the appropriate feature encoding methods, we evaluated and compared the performance of nine machine learning methods, including SVM, KNN, MLP, LR, RF, XGBoost, LightGBM, CatBoost and ensemble methods, for each feature encoding based on a randomized fivefold cross-validation. The results of classifiers based on the fivefold cross-validation are shown in Table 1. We used five feature encoding methods (AAC, PAAC, PSSM_AC, RPSSM, SSA) to convert each protein into a feature vector. As the most forceful one in five feature encoding methods, RPSSM achieved the highest AUC value in eight classifiers (Fig. 1). An interesting phenomenon is that the RPSSM feature obtained the best performance among five single features and the performance of PSSM_AC is second only to RPSSM. The evolutionary features derived from the PSSM files showed that evolutionary features have an outstanding contribution to Acrs prediction. The evolutionary feature RPSSM had a better performance than the evolutionary feature PSSM AC in most classifiers (except LR). The pre-trained machine learning feature SSA also achieved good performance for most classifiers, and its performance is better than sequence features AAC and PAAC. The PAAC contains more sequence information, showing higher AUC values than AAC for all classifiers. The sequence features AAC and PAAC achieved a relatively poor performance compared with other features. One explanation is that evolutionary features and the pre-trained feature encoded more valuable and appropriate information about protein sequences. In contrast, sequence features might involve redundant information that reduces the accuracy of Acrs prediction. In the PreAcrs model, features PAAC_AC, RPSSM and SSA were considered. From Additional file 2: Table S2, the RPSSM-based model achieved the best prediction performance among the three features on the independent test, the PSSM_AC-based model achieved the second prediction accuracy, and the SSA-based model showed a lower prediction accuracy compared to another two features. In addition, the AUC value of the PSSM_AC&SSA was 0.953, up to 0.969 after considering the feature RPSSM. Two ensemble features PSSM_AC&RPSSM and RPSSM&SSA achieved an excellent performance in terms of AUC (0.967 and 0.961, respectively). Therefore, the feature RPSSM made the most contribution to the PreAcrs model in predicting Acrs.
For most feature encodings, the LightGBM classifier, CatBoost and SVM classifier outperformed the other single classifiers (except the ensemble classifier) in terms of PRE (Table 1). This observation is supported by Fernandez-Delgado et al. [21], who found the SVM model is most likely the best classifier compared with the other 17 machine learning methods based on various public data sets. Moreover, Ke et al. [22] demonstrated LightGBM model achieved a better performance than others in multiple public datasets. LightGBM could handle the high-dimension features and large-scale data [22]. CatBoost is proved superior to XGB and LightGBM in terms of a set of publicly available datasets [23]. Although LightGBM obtained the highest PRE values among the eight classifiers in PSSM_AC and SSA in this study, CatBoost had a better performance than LightGBM in RPSSM. In addition, Catboost showed excellent performance in other metrics, such as AUC and MCC. SVM obtained the highest PRE values among the eight classifiers in features AAC and PAAC. It implied that the SVM, LightGBM and CatBoost classifiers provided an outstanding prediction ability, and SVM tended to show excellent performances in sequence features. Additionally, the highest PRE value of 1.00 was obtained by LightGBM classifier when the PSSM_AC feature was used for training during experiments. It means that the predicted positive samples of this model are more likely to be true positive samples, and it might be beneficial for the virtual screening of Acrs. To fairly compare the performance of various classifiers, other measurements were considered, such as SP, SN, and MCC. As one crucial evaluation matrix, MCC considers all four confusion matrices and can comprehensively reflect the performance. CatBoost presented its powerful and stable ability in terms of MCC value among five features. MLP outperformed other single classifiers in RPSSM features according to the MCC value. In all cases, the highest MCC value was 0.763 when the RPSSM feature was used for training in MLP. It provided more extensive and persuasive evidence for various performances with various features and classifiers. It is unreliable only to use one feature and a single model to identify Acrs protein. Although some single classifiers have shown good performance for predicting Acrs, only one classifier might not be robust and reliable enough. In order to build a more comprehensive, reliable, and robust predictor, three ensemble methods have been adopted based on eight single classifiers in this study. Three ensemble methods integrated other classifiers by three different principles. Table 1 and Fig. 2 illustrate that three ensemble methods achieved better performance than single classifiers in terms of AUC value in most features, demonstrating the superiority of ensemble learning. This observation is supported by the study of Zou et al. [24].
As we mentioned above, five features were trained by eight different classifiers, respectively. Since single features cannot comprehensively represent the Acrs for identification, we attempted to integrate five single features in two ways: ensemble feature and combination feature. For combination features, we combined singles features into a vector to train models [25–27]. We explored the contribution of a variety of combined features to the prediction models of Acrs (Additional file 1: Table S1). For ensemble features, first, we trained eight different classifiers (including ensemble classifier) with five single features, then integrated classifiers of five features as an ensemble model. This study discussed ensemble features detailly because they showed better performance than combination features. For every single feature in each classifier, we have obtained its probability score of Acrs. The output of two-feature ensemble models is obtained by averaging the predictive scores of two single features in the same model. For example, we averaged the predictive scores of predicted Acrs obtained by the AAC feature trained in the SVM model and the PAAC feature trained in the same model, and we labeled it as ‘AAC&PAAC’. Therefore, the three-feature ensemble models were obtained by averaging the predictive scores of three single features in the same model, and Feature1&Feature2&Feature3 represented the three-ensemble features. The four-ensemble features and the five-ensemble feature were also shown similarly. Finally, we used the averaged predictive scores as the final scores of the ensemble feature in every classifier. From the cross-validation results, the ensemble features achieved good performance for Acrs identification. By comparing the performance of all ensemble features, the ensemble feature PSSM_AC&RPSSM&SSA showed the best performance with the highest AUC value. The second-best ensemble feature is PAAC&PSSM_AC&RPSSM, and the PSSM_AC&RPSSM ensemble feature is the third best. We found that all the top 12 ensemble features include the RPSSM encoding method from Additional file 2: Table S2. These observations also demonstrated that the RPSSM feature plays an essential role in Acrs prediction.
In the above section, ensemble classifiers with five single features have shown an excellent ability to predict Acrs, and the Sta-LR method obtained the best performance in terms of metrics. Therefore, we used the Sta-LR classifier to train various features in this study. Besides, we compared combination features with ensemble features in the same model. The ensemble feature achieved superior performance than combination features in most classifiers. Among all models, the average AUC value of Sta-LR classifiers using PSSM_AC, RPSSM and SSA features (the three-ensemble feature PSSM_AC&RPSSM&SSA) achieved the highest 0.969. Besides, the Sta-LR classifier with PSSM_AC&RPSSM&SSA ensemble feature achieved an excellent performance in terms of a high PRE value of 0.978, a high MCC value of 0.754, an ACC value of 0.866 and an F-score of 0.848 based on the fivefold cross-validation test. Based on these findings, we constructed a PreAcrs predictor to predict Acrs with a default setting: eight machine learning classifiers (SVM, KNN, MLP, LR, RF, XGBoost, LightGBM, CatBoost) were integrated into an ensemble classifier (Sta-LR); three features PAAC_AC, RPSSM, and SSA were trained by the Sta-LR classifier, separately, and three models could be obtained in this step. Then, we could obtain the PreAcrs predictor by averaging the score of the three models. The PreAcrs predictor achieved a stable and accurate prediction performance in the fivefold cross-validation and independent dataset.
In order to further evaluate the performance of the PreAcrs predictor, we compared PreAcrs with the state-of-the-art Acrs predictor PaCRISPR. This machine learning model was proposed by Wang et al. [19], and significantly outperformed other methods such as AcRanker and BLAST on their independent dataset. Four evolutionary features, PSSM-composition, DPC PSSM, PSSM_AC and RPSSM, were adopted in the PaCRISPR predictor, which was constructed by 10 SVM classifiers. Besides, the BLAST-based predictor, AcRanker and the hidden Markov model (HMM) based predictor were implemented for the comparison. For the BLAST-based predictor, each protein in the independent dataset was searched against all samples in the training dataset based on BLAST + software [28] and was predicted as Acr when it has the highest similarity with positive samples. The predicted results of the other three predictors could be obtained from the webserver (https://pacrispr.erc.monash.edu/AcrHub). Figures 3 show that the performance of PreAcrs is better than the other predictors on the independent dataset based on the AUC and AUPRC values. The performance demonstrates that the PreAcrs method is more suitable for capturing the intrinsic patterns of non-homologous Acrs than other predictors. From other metrics (Table 2), HMM obtained higher PRE and SP values than PreAcrs, but it does not indicate that HMM outperformed PreAcrs. It means the false positive is lower and one possible reason for it is HMM prone to predict the queried proteins as non-Acrs. HMM uses probabilistic models to search homologous protein sequences. The homology-based baseline predictors made a biased prediction, as HMM failed to recognize Acrs. It predicted the Acrs with extremely high accuracy (the lowest FP) but classified many true Acrs into non-Acrs (the highest FN). HMM obtained the best PRE with the cost of predicting most Arcs as non-Acrs. This observation is supported by the work of Wang et al. [19]. Therefore, when considering the FN and FP, HMM showed poor performance when it was evaluated. According to other more critical metrics like ACC, F-score and MCC, PreAcrs outperformed the other four approaches. We listed the predictive scores of five experimentally validated Acrs on the independent test as a case study to further evaluate the performance of PreAcrs (Table 3). The PreAcrs achieved better performance than PaCRISPR and AcRanker. For the AcrIIA7 and AcrIIA9, PaCRISPR predicted lower scores, and the predictive score of AcrIIA7 was 0.407. In contrast, PreAcrs gave these three Acrs higher scores. For AcrIIC2, PaCRISPR showed better performance, but PreAcrs also gave considerable scores. PaCRISPR only considered four features driven from evolution information and the SVM model, while PreAcrs incorporated the SSA feature from the pre-trained model and eight different models. Considering more information and various classifiers, PreAcrs showed a more robust and accurate prediction performance.
The identification of candidate Acrs plays a vital role in manipulating CRISPR-Cas machinery as a tool in gene editing or gene therapy. Using the machine learning method to identify the new Acrs based on the protein sequence can accelerate the discovery of Acrs. In this work, we proposed a machine learning-based ensemble framework, PreAcrs, to accurately and efficiently identify Acrs from protein sequences. PreAcrs extracted distinctive characteristics from experimentally validated Acrs by combining the evolutionary features with the pretrained model feature with multiple models. The features were trained by an ensemble classifier constructed by eight base classifiers. PreAcrs predictor displayed a good performance for predicting new Acrs in terms of prediction accuracy and robustness. We anticipate that PreAcrs will be extensively used in Acrs prediction and help researchers to have a comprehension understanding of Acrs. PreAcrs shows excellent performance compared to the existing methods, but it still has some limitations. One limitation is that only the mRMR algorithm is applied to select significant features in PreAcrs, so some biases in this step may reduce the predictive accuracy. Another limitation is that PreAcrs does not provide a visual and user-friendly website; it may be difficult for some biologists to analyze Acrs. In future works, we may use multiple feature selection algorithms to calculate feature importance to obtain a reasonable feature, and build a powerful, user-friendly and interactive website.
Figure 4 shows the overall workflow of the PreAcrs framework, including five major steps: Dataset collection and curation, Feature encoding, Feature selection, Model training, and Model validation. These steps are described in the following sections.
To build a powerful Acrs predictive model, we need to construct a training dataset and an independent test dataset comprised of two parts: positive samples (experimentally validated Acrs) and negative samples (non-Acrs). As mentioned above, Anti-CRISPRdb, AcrDB, and AcrCatalog are online databases of anti-CRISPR proteins. The latest update time of the Anti-CRISPRdb database is January 2021, and it has 1378 experimentally validated entries. The AcrDB and AcrCatalog are databases of computationally predicted Acrs. In this study, we collected the experimentally validated Acrs from Anti-CRISPRdb, which is the latest database and contains more experimentally validated Acrs than others. We extracted 1,378 experimentally validated Acrs from the Anti-CRISPRdb [12] and 17 newly discovered experimentally validated Acrs from NCBI. To construct a robust machine learning model and eliminate the redundant Arcs, we used CD-HIT [29] to remove the highly-homologous sequences. Here, we set the identification threshold as 70% in CD-HIT (removed those sequences with more than 70% similarity). 588 Acrs sequences were obtained, and their length ranges from 50 to 350. After the 588 Acrs were randomly divided into two parts with a ratio of 7:3, we obtained 412 Acrs in the training dataset and 176 Acrs in the independent dataset. Because there is no standard set of non-Arcs, constructing a comprehensive and reasonable non-Acrs dataset is a challenging and vital question. In this study, we referred to the work of Wang et al. [19] to construct the non-Acrs dataset. Because the range of Acrs sequence length is fixed, and most Acrs were found from a limited set of phages and mobile genetic elements (MGEs), the negative samples were selected with four strict criteria from Uniprot. The four criteria are the following: (1) must not be known or putative Acrs; (2) must be isolated from phage or bacterial MGEs (known or putative MEGs); (3) must have < 40% sequence similarity to each other and the 588 positive samples; (4) the lengths must fall in the range between 50 and 350 residues. According to the above four criteria, 1571 non-Acrs were obtained in this study. Then, we randomly selected 412 non-Acrs as negative samples in the training dataset and 176 non-Acrs as negative samples in the independent dataset. Each negative sample was only included in one dataset. In this way, the training dataset has 412 positive and 412 negative samples, while the independent test dataset contains 176 positive and 176 negative samples (Table 4). In addition, we chose 5 Acrs from the independent dataset as a case study.
In order to find the features that could better represent Acrs, we firstly evaluated 18 types of features to represent Acrs, including the composition of k-spaced amino acid pairs (CKSAAP), amino acid composition (AAC), pseudo amino acid composition (PAAC), bidirectional long short-term memory (BiLSTM), soft sequence alignment (SSA), PSSM_AC, RPSSM and PSSM-composition et. (Table 5 and Additional file 3: Table S3). We selected five features (AAC, PAAC, PSSM_AC, RPSSM, SSA) considering the computational requirements and predictive performance. The five features could be categorized into three groups: sequence features, evolutionary features, and pre-trained model features. These features have been widely applied in feature encoding research [19, 30, 31] and have achieved a good performance in protein properties and function predictions [32–38]. The following are the five features adopted in this study.
As one of the most important features, amino acid composition (AAC) has been successfully applied in many bioinformatics fields, for example, protein structure classification [30], thermophilic proteins prediction [39], and protein–protein interactions identification [40]. For AAC, each sequence is represented by a 20-dimensional numerical vector, in which each number corresponds to the frequency of an amino acid type in the whole protein sequence [41]. Every element in AAC of a given protein could be calculated by the following formula:withwhere is the number of type native amino acid in the whole protein sequence, and is the length of the protein sequence. Finally, the is the frequency of type native amino acid in the protein .
Pseudo-Amino acid composition (PAAC) was proposed by Zhou [42] for predicting cellular protein attributes and has been widely used in many studies [31, 43]. This group of descriptors involves sequence-order information, hydrophobicity value, hydrophilicity value, and side-chain mass. The PAAC is defined by 20 + λ discrete numbers:withwhere the is the normalized frequency of amino acid in the protein sequence. L is the length of protein and θj is the jth rank of the coupling factor. represents the correlation function, and λ is the maximum correlation length. This study used iLearnPlus to extract PAAC feature-based protein sequences [44] and generated a 23-dimensional feature vector for each protein.
PSSM-AC is derived from Position-Specific Scoring Matrix (PSSM) by applying the auto covariance (AC) transformation to each column of PSSM, and it measures the average correlation between two elements within the PSSM [45, 46]. A 20 × G-dimensional vector represents each sequence in PSSM-AC by the following formula:withwhere represents the PSSM value at the row and jith column, and the is the average value of amino acid j in the whole protein sequence. is a number smaller than the length of the whole protein sequence L, and the ranges from 1, 2, …, G; here, is set to 10 in this study [47]. Therefore, a 200-dimensional feature vector is generated for each protein.
According to the work of Li et al. [48], the original PSSM profile (L × 20) could be reduced to a L × 10 matrix by merging some columns. RPSSM is obtained by exploring the local sequence information based on the L × 10 reduced PSSM [49, 50]:andwhere represent the 20 columns in the original PSSM profile corresponding to the 20 amino acids. The re-PSSM is further transformed into a 10-dimensional vector:and Additionally, the re-PSSM can be further transformed into a 10 × 10 matrix to capture the local sequence-order information by this formula:where represents the element at the ith row and jth column of there-PSSM. Finally, a 110-dimensional RPSSM feature is obtained by combining and :
The pretrained SSA embedding mosdel is obtained by combining the pre-trained language model with the soft sequence alignment (SSA) [51]. First, an embedding matrix RL×121 is given using the stacked BiLSTM encoders for each sequence, where L is the protein sequence length [52]. Then, the pretrained SSA embedding model is trained and optimized by SSA, which the following formulas could describe. For convenience, we supposed two embedding matrices P1(RL1×121) and P2(RL2×121), of two different protein sequences with lengths L1 and L2, respectively:where xi, yi are vectors with 121-dimension. The following formula represents the similarity of P1 and P2:andwith The SSA embedding model could convert each protein sequence into an embedded matrix RL×121, and finally, an average pooling operation obtained a 121-dimensional feature.
Original features are represented by a high dimensional vector or matrix, which would raise severe problems in machine learning algorithms, such as overfitting, time-consuming training process and high requirement of computing resources. Therefore, identifying the most contributing information and features plays a vital role in performance improvement. As one of the most popular feature selection algorithms, maximum relevance minimum redundancy (mRMR) was proposed by Peng et al. [53] and has been applied in many studies and achieved robust performances [54–56]. In this study, mRMR was used to identify the most important features and improve the generalization ability of the model.
In this study, we focused on the traditional machine learning classification methods, including support vector machine, k-nearest neighbor, multi-layer perceptron, logistic regression, random forest, extreme gradient boosting, Light gradient boost machine and ensemble method that integrates the previous eight classification methods by hard voting strategy and stacking classifiers. More information is shown in the following subsections.
Support vector machine (SVM) was first proposed by Vapnik et al. [57], and has successfully dealt with some binary classification problems in bioinformatics [25, 58, 59]. Two parameters Cost (C) and Gamma (γ) affect the performance of the SVM model with the RBF kernel. In this study, we used the grid search strategy to optimize C and γ in the space {2−6, 2−5, …, 25, 26}. Finally, an SVM classifier with the optimal value of C and γ was constructed.
K-nearest neighbor (KNN) is a fundamental classifier that has been applied in predicting protein function [60], extracting protein–protein information [61], and predicting eukaryotic protein subcellular [62]. The performance of KNN is directly affected by the parameter k. In this study, a grid search within the space was applied to optimize the parameter k during model training, where FeaNum is the number of features used in modelling.
Multi-layer perceptron (MLP) is known as a type of artificial neural network (ANN) [63, 64]. MLP has been applied in many bioinformatics studies, such as the prediction of protein structure classes [65], protein tertiary structure [66], and DNA–protein binding sites [67]. In this study, an MLP classifier with two hidden layers was trained, and the first and second hidden layers have 64 and 32 nodes, respectively. The maximum learning iterations is 1000.
Logistic regression (LR) is widely used to predict the probability of an event happening [59, 68], which the following formula could represent:where p(y) is the expected probability of dependent variable , and β0 and β1 are constants.
Random forest (RF) classifier is proposed by Breiman [69] and has been used in the prediction of type IV secreted effector proteins [70] and protein structural class [59]. To find the optimal number of the trees M and features mtry, we used a gird searching to optimize and within space and {1, 6, 11, 16}, respectively, where FeaNum is the number of features adopted during modeling.
Extreme gradient boosting (XGBoost) is a scalable end-to-end tree boosting system [71] and has been widely used as a fast and highly effective machine learning method [72, 73]. Eitzinger et al. implemented AcRanker using XGBoost to identify Acrs [14, 16]. In this study, the default parameters are adopted in the XGBoost model, except for the learning rate of 0.1.
Light gradient boost machine (LightGBM) shows excellent performance when the feature dimension is high and the larger data size [21]. LightGBM has been used in identifying miRNA targets [74] and predicting the protein–protein interactions [75] and the blood–brain-barrier penetration [76]. This study used the LightGBM package with default parameters in python during experiments.
CatBoost achieves state-of-the-art results since it successfully handles categorical features and calculates leaf values via a new scheme, which helps reduce overfitting [23]. Catboost has been applied in various tasks, including molecular structure relationship and the biological activity prediction [77] and the identification of pyroptosis-related molecular subtypes of lung adenocarcinoma [78]. In this study, the parameters of CatBoost were set as default values.
This study proposed three ensemble models to construct more robust and reliable classifiers, which predicted new Acrs proteins by integrating the above eight classifiers (SVM, KNN, MLP, LR, RF, XGB, LightGBM, and CatBoost) through the hard voting rule (Ens-vote) or two stacking classifiers with logistic regression (Sta-LR) and gradient boosting classifier (Sta-GBC) [79], respectively.
Fairly evaluating the classification methods' predictive performance is an essential subject in machine learning. In this study, we used six measurements, namely, Sensitivity (SN), Specificity (SP), Accuracy (ACC), Precision (PRE), F1-score, and Matthew’s correlation coefficient (MCC) [80], which are denoted as:where TP, TN, FP, and FN are the number of true positive, true negative, false positive and false negative, respectively. Besides, the area under the receiver operating characteristic (ROC) curve (AUC) is also used to assess the performance, and the ROC was shown in a plot of the TP rate versus the FP rate. All methods were evaluated based on a fivefold cross-validation.
Additional file 1: Table S1. Performance of all single features.Additional file 1: Table S2. Performance of ensemble features.Additional file 1: Table S3. Performance of combinational features. | true | true | true |
|
PMC9598032 | Hyueyun Kim,Ji Ha Choi,Chang Mo Moon,Jihee Lee Kang,Minna Woo,Minsuk Kim | Shrimp miR-965 transfers tumoricidal mitochondria | 26-10-2022 | Mitochondria,Microfluidics,Tunneling nanotube,Breast tumor,Cotton candy,Tomographic microscope,Refractive index | Background Micro RNA of Marsupenaeus japonicas has been known to promote apoptosis of tumor cells. However, the detailed mechanisms are not well understood. Results Using tomographic microscope, which can detect the internal structure of cells, we observed breast tumor cells following treatment of the miRNA. Intriguingly, we found that mitochondria migrate to an adjacent tumor cells through a tunneling nanotube. To recapitulate this process, we engineered a microfluidic device through which mitochondria were transferred. We show that this mitochondrial transfer process released endonuclease G (Endo G) into tumor cells, which we referred to herein as unsealed mitochondria. Importantly, Endo G depleted mitochondria alone did not have tumoricidal effects. Moreover, unsealed mitochondria had synergistic apoptotic effects with subtoxic dose of doxorubicin thereby mitigating cardiotoxicity. Conclusions Together, we show that the mitochondrial transfer through microfluidics can provide potential novel strategies towards tumor cell death. Supplementary Information The online version contains supplementary material available at 10.1186/s12575-022-00178-8. | Shrimp miR-965 transfers tumoricidal mitochondria
Micro RNA of Marsupenaeus japonicas has been known to promote apoptosis of tumor cells. However, the detailed mechanisms are not well understood.
Using tomographic microscope, which can detect the internal structure of cells, we observed breast tumor cells following treatment of the miRNA. Intriguingly, we found that mitochondria migrate to an adjacent tumor cells through a tunneling nanotube. To recapitulate this process, we engineered a microfluidic device through which mitochondria were transferred. We show that this mitochondrial transfer process released endonuclease G (Endo G) into tumor cells, which we referred to herein as unsealed mitochondria. Importantly, Endo G depleted mitochondria alone did not have tumoricidal effects. Moreover, unsealed mitochondria had synergistic apoptotic effects with subtoxic dose of doxorubicin thereby mitigating cardiotoxicity.
Together, we show that the mitochondrial transfer through microfluidics can provide potential novel strategies towards tumor cell death.
The online version contains supplementary material available at 10.1186/s12575-022-00178-8.
Tunneling nanotubes are thin and long membrane elongations between cells that mediate trafficking of subcellular vesicles, proteins, and organelles [1]. They consist of filamentous-actin and are about 0.05–1 µm in width and 100 µm in length [1] which has been implicated as one mechanism by which mitochondria can be transferred from one cell to another [2]. The speed of mitochondrial migration is about 80–1400 nm/s through tunneling nanotubes [3]. Several studies have been conducted on mitochondrial migration between cells [4–7]. An in vitro study showed that oxidative stress-induced apoptosis of endothelial or H9c2 cardiomyocytes was abolished by transferring of mitochondria from mesenchymal stem cells [8, 9]. Another in vivo study showed that mitochondrial transfer inhibited the hypoxia-based apoptosis in cardiomyocytes [10]. Recent studies suggest that mitochondria can be transferred to support the survival of metabolically compromised cells [11, 12]. Moreover, it has been reported that cancer cells can hijack mitochondria from immune cells via physical nanotubes [13]. Therefore, mitochondrial transfer through tunneling nanotubes can provide important clues in our understanding of tumor cell fate. We describe here the application of optical tomography whereby a remote imaging technique is used to obtain cross-sections of cells [14]. In optical tomography, projected images are obtained by waves passing through the cells at various angles, and digital images are subsequently reconstructed to obtain the internal structure of cells in a cross-sectional manner [14, 15]. Since the tomographic microscope can reconstruct cell images according to varying refractive index, transparent objects such as lipid droplets and mitochondria can be detected without staining [16, 17]. In this study, we found extracellular vesicles and microRNAs (miRNA) from shell membrane of Marsupenaeus japonicas. A miRNA is a small non-coding RNA molecule containing about 22 nucleotides that mediates RNA silencing and posttranscriptional regulation of gene expression [18, 19]. A miRNA can be released into extracellular space and taken up by neighboring cells [20]. From miRNA profiling studies, multiple miRNAs, including miR-6491, miR-6492, miR-6493, miR-6494, or miR-965 have been reported from Marsupenaeus japonicas [21, 22]. One of these, miR-965 has been shown to decrease the expression of myeloid cell leukemia-1 (Mcl-1), which blocks the secretion of cytochrome C from mitochondria [22–24]. Here, we examined breast tumor cells and their response to miR-965 and its antagomir by tomographic microscopy. Intriguingly, we found that mitochondria within tumor cells transferred into neighboring tumor cells via tunneling nanotubes and induced apoptosis. Therefore, we hypothesized that the mitochondria of tunneling nanotubes can be effective novel strategy towards tumor cell death. To this end, we conducted experiments on the mitochondrial mechanisms of tumor apoptosis following miR-965 treatment.
The shell membrane of Marsupenaeus japonicas (Fig. 1A), contained a number of microvesicles which were readily visible by electron microscopy (Fig. 1B). In contrast, no microvesicles were visible in the isolated muscle (Fig. 1C), through various miRNAs were measured except for mja-miR-965 (Fig. 1D). Intriguingly, mja-miR-965 constituted the predominant fraction in the shell membrane (Fig. 1E). As the shell of crustaceans or miR-965 have been shown to inhibit tumor growth [22, 24–26], we tested the effects of synthetic miR-965 on MDA-MB-453 breast cancer cells and found this to increase apoptosis to 10% of tumor cells compared to less than 0.8% in control group as assessed by ELISA to detect ssDNA (Fig. 1F). Similarly, miR-965 increased cleaved caspase-3, cytochrome C and Endo G (Fig. 1G-J), whereas Mcl-1 levels were decelerated (Fig. 1K). To further visualize these MDA-MB-453 cells, they were scanned with a λ = 520 nm laser beam of tomographic microscopy and the holographic images were recorded and shown as 3D rendering (Fig. 2A and B, Supplementary Fig. 1A-C). Next, we exposed miR-965 or antagomir-965 (antisense of miR-965) to the tumor cells. For easy identification, adjacent cells were digitally colored in red or green and named R and G. After 4 h of exposure with miR-965, G cells were detached and suspended in cell media, which showed morphological alterations such as cell shrinkage and membrane blebbing (Fig. 2A). However, these morphological changes did not occur following treatment with the antagomir (Fig. 2B). The miR-965 again increased apoptosis, as showing by increased cleaved caspase-3, cytochrome C, and Endo G (Fig. 2C-G) with decreased levels of Mcl-1 (Fig. 2H) and its antagomir attenuated the miRNA-induced apoptotic changes.
To examine the cellular properties in more depth, refractive index images were magnified using tomographic microscopy. Intriguingly, we detected tubular-shaped parts moving from one cell to another and mitotracker staining confirmed that these were mitochondria (Fig. 3A). Migration of mitochondria between cells has been reported to improve the viability or metabolism of the recipient cell [2, 9, 12]. We hypothesized that mitochondrial migration induces apoptosis in the recipient cell. To determine this, MDA-MB-453 cells were treated with miRNA negative control (NC), mja-miR-965, or antagomir-965 (Fig. 3B). In addition, we designed a microfluidic system to create an environment similar to nanotubes (Fig. 3B, Supplementary Fig. 1D). Considering that the diameter of the nanotube is 10–1000 nm and the mitochondrial diameter is approximately 500 nm, we devised a microfluidic system using polydimethylsiloxane (PDMS) which was poured into cotton candy to make a mold followed by addition of water to remove to remove the filaments. The microfluidic device was scanned with an electron microscope and the diameter measured to be about 950 nm (Fig. 3C). The isolated mitochondria were introduced into the microfluidic device at a flow rate of 10–30 μm/s. The mitochondria were stained with mitotracker which were shown to remain detectable up to 10 h in recipient tumor cells (Fig. 3D). Mitochondria that have passed through the microfluidic device are referred to herein as unsealed mitochondria (Fig. 3B). We next treated MDA-MB-453 cells with either the isolated or unsealed mitochondria. The isolated mitochondria or unsealed mitochondria did not lead to an increase in apoptosis, caspase-3, or cytochrome C in recipient cells (Fig. 3E-H). However, unsealed mitochondria extracted from miR-965-treated cells led to an increase in apoptosis of the recipient cells (Fig. 3E-H). Furthermore, regardless of miR-965 or antagomir treatment, unsealed mitochondria increased Endo G expression in recipient cells (Fig. 3I). Also both isolated and unsealed mitochondria extracted from miR-965-treated cells decreased Mcl-1 in recipient cells (Fig. 3J). Finally, antagomir treatment attenuated the miRNA-induced apoptotic genes (Fig. 3E-H), with the exception of Endo G (Fig. 3I).
We noted an accumulation of Endo G protein in the subcellular region proximal to the adjacent cell at 66 min post treatment with miR-965 (Fig. 4A, Supplementary Fig. 2A). This surge in protein levels was not associated with any differences in mRNA levels (Fig. 4B, Supplementary Fig. 2B-H). We posit that Endo G was released by the tunneled mitochondria. In order to better understand the effects of mitochondrial Endo G, ENDOG was knocked down using CRISPR-plasmid and confirmed using a real time-PCR (Fig. 4C). Following treatment with miR-965, mitochondria were isolated from MDA-MB-453 cells with or without ENDOG knockdown (Fig. 4D). ENDOG deficient cells were then treated with unsealed mitochondria referred herein as unsealed mitochondria β. Regardless of microRNA treatment, the unsealed mitochondria β from ENDOG deficient cells did not induce apoptosis (Fig. 4E). However, the unsealed mitochondria β from ENDOG deficient cells still increased caspase-3 and cytochrome C (Fig. 4F-H). In addition, the unsealed mitochondria β did not increase Endo G (Fig. 4I), but decreased Mcl-1 levels (Fig. 4J). Taken together the results indicate that disgorged Endo G through tunneling nanotubes is essential to promote tumor cell apoptosis.
In order to assess the lethal effects of unsealed mitochondria in vivo, mice with DMBA-induced mammary carcinoma were treated with miR-965, doxorubicin, or unsealed mitochondria (Fig. 5A). We refer to isolated mitochondria from the miR-965 treated cells that have passed through the microfluidics as unsealed mitochondria γ (Fig. 5B). 5 mg/kg of doxorubicin reduced the tumor volume, whereas 0.05 mg/kg of doxorubicin or miR-965 alone did not have any tumoritoxic effects (Fig. 5C). However, even at low concentrations of doxorubicin at 0.05 mg/kg, tumor growth was inhibited together when combined with unsealed mitochondria γ (Fig. 5C and D). In keeping with well-known cardiotoxic effects of doxorubicin, cardiac fractional shortening and ejection fraction decreased in the doxorubicin group at 5 mg/kg (Fig. 5E-G). However, 0.05 mg/kg of doxorubicin or the unsealed mitochondria γ did not lead to any adverse cardiac effects (Supplementary Fig. 1E and F). Overall, mitochondria through tunneling nanotubes can effectively synergize with subtoxic levels of doxorubicin to cause tumor cell death (Fig. 5H).
As the shell of crustaceans has been known to inhibit the growth of tumor, we carefully observed the shell structure of Marsupenaeus japonicas [25, 26]. In the shell membrane, a number of micro vesicles were found with predominance of mja-miR-965. We found that miR-965 was effective in inducing apoptosis of MDA-MB-453 cancer cells through transcellular migration of mitochondria via tunneling nanotubes. While transcellular migration of mitochondria through tunneling nanotubes has been generally shown to impede apoptosis and optimize metabolism [3, 6, 9], we show in this study that this notion can also cause cytotoxic effects. Biggest challenge in this area of investigation has been with obtaining accurate imaging of tunneling nanotubes and mitochondria using optical microscopy. Difficulties around processing using chemicals and fixatives in addition to the limitations of two-dimensionality can contribute to the overall limitations of accurate assessment of tunneling nanotubes and mitochondria. Therefore we used tomographic microscopy which can provide the internal structure of cell images from multiple angles and render digitally reconstructed images according to the difference in refractive index, whereby tunneling nanotubes and mitochondria can be distinguished and visualized without special staining [14, 16]. Using such tomographic images, we reveal that mitochondria migrated to an adjacent cell through tunneling nanotubes to trigger apoptosis. Intriguingly, migration of mitochondria alone did not lead to apoptosis. Therefore, we hypothesized that an additional factor was needed for apoptosis to proceed. It has been known that in addition to caspase-3, Endo G is critical for apoptosis [27]. Intriguingly, we found an increase in Endo G following passing of mitochondria through the tunneling nanotube. Since this was not associated with mRNA of Endo G, we posit that this accumulation of Endo G occurred as a result of release from transported mitochondria. To recapitulate this process, PDMS was poured onto cotton candy fibers with a diameter of about 950 nm to create a tunneling nanotube-like device. Mitochondria that have passed through the narrow nano-sized area which we referred to herein as unsealed mitochondria and the organelles from miR-965-treated cells induced apoptosis of breast tumor cells. In contrast, mitochondria of cells treated with its antagomir did not induce the tumor cell death. Moreover, unsealed mitochondria increased the amount of Endo G in recipient cancer cells regardless of miR-965. To confirm the relationship between mitochondria and Endo G, experiments were performed on breast tumor cells deficient in ENDOG. Mitochondrial delivery from miR-965-treated ENDOG-deficient cells did not induce cell death despite passing through a microfluidic system. In summary, miR-965 triggers effective tumor apoptosis by releasing Endo G in the recipient cells after transcellular migration of mitochondria. To assess the role of this process in vivo, mice with DMBA-induced mammary carcinoma were administered with doxorubicin or unsealed mitochondria. While unsealed mitochondria from miR-965-alone did not have tumoricidal effects, it effectively synergized with subtoxic dose doxorubicin in reducing tumor size thereby sparing its well-known cardiotoxic effects. Although the mechanism of the unsealed mitochondria in the heart is unknown, reported studies suggest that Endo G is involved in cell survival rather than apoptosis of cardiomyocytes [28]. Taken together, we show that the mitochondria of microfluidics may provide novel strategies to effectively kill tumor cells. Lethal effects of mitochondria can potentially be harnessed using physical properties of delivery.
In summary, our study revealed that shrimp miRNA could regulate the number of breast tumor cells. The mechanism was involved in lethal mitochondria through tunneling nanotubes. We reproduced the physical properties of delivery with microfluidics, which provide novel strategies to effectively kill tumor cells.
All mice were in C57BL/6 (RRID: 2,159,769) background and purchased from Charles River. All experiments using animals were approved and performed by the Ewha Womans University Animal Care Committee (Guide for the care and use of laboratory animals; IRB number: ESM14-0260). Forty female mice were each given weekly 1 mg doses of 7,12-Dimethylbenzathracene (DMBA) in 0.2 ml of sesame oil by oral gavage for six weeks and were implanted with 30 mg pellets of compressed medroxyprogesterone acetate (MPA) subcutaneously beginning at 5 weeks of age. Doxorubicin was administered weekly by intraperitoneal injection (5 mg/kg body weight, 300 μl injection volume) for four weeks. Mice were then maintained for an additional week.
Human breast cancer cell lines, MDA-MB-453 (ATCC; RRID: CVCL_0418), was maintained in Dulbecco’s MEM (11,885, Gibco, USA) with 10% fetal bovine serum (16,000,044, Gibco, USA) at 37 °C under an atmosphere of 95% O2 and 5% CO2. Cells were exposed to miRNA negative control (MMIR-000-PA-1, System Biosciences), synthetic miR-965 or its antagomir (Bioneer, South Korea) using a kit from System Biosciences (EXFT10A-1, System Biosciences) for 5 h.
Green light (λ = 520 nm, exposure 0.2 mw/mm2) from a laser diode was split into cells and reference beam at Nanolive (3D cell explorer, Switzerland). Cells were illuminated with a laser beam inclined at 45° which rotated around the sample 360°. Holographic images were recorded on a digital camera by combining the beam that had passed through the cells with the reference beam. 3D cell images were recorded up to 30 μm depth of reconstruction. Mitochondria were visualized with MitoTracker (M7514, Thermo Fisher Scientific, USA) at 490 nm.
Cotton candy sheets were sealed with polydimethylsiloxane (PDMS) to construct microfluidic mold. After hardening, cotton candy fibers were removed by perfusing with water to make microfluidic mold of approximately 950 nm. Each microfluidic device was connected by polythene tubing (PE10, Braintree scientific, USA) with an inner diameter of 0.28 mm. Fluid flow was controlled by individual peristaltic pump (3,200,243, Dolomite, UK). Isolated mitochondria were introduced to microfluidic devices at a flow rate of 10–30 μm/s. After passing through the microfluidic channels, mitochondria were transported into cells.
RNA levels of ENDOG (Hs00172770_m1) were determined using primer/probe set from Life Technology. Real-time PCR was performed with TaqMan universal PCR Master Mix on an ABI Real time PCR System 7000 (Applied Biosystems, USA). PCR conditions were 50 °C for 2 min and 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. For each experimental sample, the relative abundance value was normalized to the value derived from ACTB (Hs03023943_g1, Life Technology) as housekeeping control gene. Relative mRNA levels were quantified using the comparative 2−ΔΔCT method. The extracted miRNA were determined using microarray service (Macrogen, South Korea).
Apoptosis was determined using ssDNA ELISA Kit (APT225, Sigma-Aldrich, USA). Briefly, cell plates were fixed and incubated with ABTS solution for 30 min to allow binding to HRP at 37 °C. To denature DNA, cell plates were incubated for 20 min at 75 °C. After cooling at 4 °C for 5 min, plates were blocked in 5% skim milk (70,166, Sigma-Aldrich) in PBS at 37 °C for 1 h. Cells were incubated with antisera mixture for 30 min and washed with PBS. After treatment with stop solution, absorbance was measured at 405 nm. Caspase-3 activity was determined using Cleaved Caspase-3 ELISA kit (ab220655, abcam, USA). Cells were fixed and incubated with antibody cocktail at 37 °C for 1 h, then washed and incubated with TMB solution for 30 min. After treatment with stop solution, absorbance was measured at 450 nm.
Cells were homogenized and centrifuged at 5,000 g for 20 min. Protein content of the supernatant was diluted, boiled with sample loading dye, and 100 mg were loaded in SDS-PAGE (4561033EDU, Bio-Rad). After blotting, membranes were blocked in 5% skim milk (70,166, Sigma-Aldrich) in PBS containing 0.1% Tween-20 (P1379, Sigma-Aldrich). Membranes were incubated with antisera directed against cytochrome C (1:1000; #11,940, RRID: AB_2637071, Cell signaling technology, USA), Endo G (1:1000; #4969, RRID: AB_2098768, Cell signaling technology, USA), Mcl-1 (1:1000; ab28147, RRID: AB_776246, abcam, USA), or β-actin (1:1000; sc-47778, RRID: AB_626632, Santacruz Biotechnology, USA), then with secondary antibodies (mouse-specific HRP-conjugated antibody or rabbit-specific HRP-conjugated antibody). Bands were visualized using ECL (32,106, Thermo Scientific) detection kit and quantified by densitometry.
Cells were harvested from culture dishes with homogenization buffer (20 mM HEPES–KOH, 220 mM mannitol, and 70 mM sucrose) containing a protease inhibitor mixture (Sigma-Aldrich) and centrifuged at 2300 × g for 5 min. The cell pellet was resuspended with homogenization buffer and incubated on ice for 5 min at 4 °C. Cells were ruptured by 10 strokes using a 27-gauge needle. The homogenate was centrifuged at 5800 × g for 5 min, and mitochondria were harvested. The amount of isolated mitochondria was expressed as protein concentration by using the Bio-Rad protein assay kit (Bio-Rad, Richmond, USA). Mitochondrial transfer was conducted by co-incubating isolated mitochondria with cells (1 × 105 cells/well of a 6-well plate) at 37 °C under 5% CO2 for up to 10 h. For in vivo experiments, mice were administered weekly with 1 × 104 isolated mitochondria per gram of body weight via the tail vein.
We fixed the tissues of Marsupenaeus japonicas or mold of cotton candy-microfluidic with 3% buffered glutaraldehyde (G5882, Sigma-Aldrich) for 2 h and processed into resin (02,334, Polysciences, German). After embedding, the resin block was thin-sectioned by ultramicrotomy. Sections of 50–70 nm thickness were collected on metal mesh and stained with electron dense particles before imaging of ultrastructures, using the transmission electron microscope (H-7650, Hitachi-Science & Technology, Japan).
Clustered regularly interspaced short palindromic repeats (CRISPR) transfection of ENDOG in MDA-MB-453 was performed using a kit from Santa Cruz (sc-395739, Santacruz Biotechnology, USA). Briefly, in six-well culture plates, 106 cells were plated and exposed to the ENDOG plasmid (sc-403263, Santacruz Biotechnology, USA) or negative control-CRISPR plasmid (sc-418922, Santacruz Biotechnology, USA) solution for 8 h at 37 ℃ in a CO2 incubator. Then, media was changed to Dulbecco’s MEM with 10% fetal bovine serum and incubated for another 18 h. The ENDOG expression was determined using RT-PCR.
For echocardiography, Vevo 2100 was used at Cardiovascular Research Center in Seoul. Mice were anesthetized with 2% isoflurane and maintained with 1.5% isoflurane followed by application of depilatory cream to the chest and wiped clean to remove all hair in the area of interest. The scanning probe (20 MHz) was used to obtain 2D images of the parasternal long axis. These 2D images were converted to M-mode.
Values were means ± SE. The significance of differences was determined by a two-way analysis of variance (ANOVA), or a one way ANOVA followed by a Bonferroni post-hoc analysis where appropriate. Differences were considered significant when P < 0.05.
Additional file 1: Supplementary Figure 1. Morphology and apoptosis in breast tumor cells. (A) Using tomographic microscopy, the number of filopodia was measured in primary epithelial breast tumor cells and MDA-MB-453. (B and C) Viability and morphology of MDA-MB-453 were observed for 48 h. (D) Efficiency of stained mitochondria uptake into MDA-MB-453. (E) Comparison of unsealed mitochondria derived from DMBA-induced mammary carcinoma. (F) Role of unsealed mitochondria on apoptosis in cardiomyocytes and breast epithelial cells. Results are the means ± SE of 6 experiments in each group. *Significantly different from treatment of Rapamycin and Y27632 for 0 h, P < 0.05. #Significantly different from unsealed mitochondria on isolated cardiomyocyte, P < 0.05.Additiona file 2: Supplementary Figure 2. The relationship between Endo G and apoptosis. (A) Following treatment of miR-965, refractive index images of MDA-MB-453 were immunostained to detect Endo G (orange) or mitochondria (green). Adjacent cells were digitally colored in red or green and named R and G. (B-E) Isolated mitochondria or unsealed mitochondria were produced in large quantities from approximately 6 X 108 MDA-MB-453 cells. After precipitating mitochondria with a centrifuge, the supernatant of these samples were lyophilized and western blotting was performed for Endo G detection. (E-H) Transfection of Endo G ORF lentiviral particles (Origene, RC205089L1V) into MDA-MB-453 cells did not induce apoptosis. Results are the means ± SE of 6 experiments in each group. *Significantly different from treatment of isolated mitochondria, P < 0.05. #Significantly different from treatment of miRNA negative control, P < 0.05. | true | true | true |
|
PMC9598034 | Tian Du,Lu Pan,Chengyou Zheng,Keming Chen,Yuanzhong Yang,Jiewei Chen,Xue Chao,Mei Li,Jiabin Lu,Rongzhen Luo,Jinhui Zhang,Yu Wu,Jiehua He,Dongping Jiang,Peng Sun | Matrix Gla protein (MGP), GATA3, and TRPS1: a novel diagnostic panel to determine breast origin | 25-10-2022 | Breast carcinoma,Immunohistochemical,MGP,TRPS1,GATA3 | Background Metastatic breast carcinoma is commonly considered during differential diagnosis when metastatic disease is detected in females. In addition to the tumor morphology and documented clinical history, sensitive and specific immunohistochemical (IHC) markers such as GCDFP-15, mammaglobin, and GATA3 are helpful for determining breast origin. However, these markers are reported to show lower sensitivity in certain subtypes, such as triple-negative breast cancer (TNBC). Materials and methods Using bioinformatics analyses, we identified a potential diagnostic panel to determine breast origin: matrix Gla protein (MGP), transcriptional repressor GATA binding 1 (TRPS1), and GATA-binding protein 3 (GATA3). We compared MGP, TRPS1, and GATA3 expression in different subtypes of breast carcinoma of (n = 1201) using IHC. As a newly identified marker, MGP expression was also evaluated in solid tumors (n = 2384) and normal tissues (n = 1351) from different organs. Results MGP and TRPS1 had comparable positive expression in HER2-positive (91.2% vs. 92.0%, p = 0.79) and TNBC subtypes (87.3% vs. 91.2%, p = 0.18). GATA3 expression was lower than MGP (p < 0.001) or TRPS1 (p < 0.001), especially in HER2-positive (77.0%, p < 0.001) and TNBC (43.3%, p < 0.001) subtypes. TRPS1 had the highest positivity rate (97.9%) in metaplastic TNBCs, followed by MGP (88.6%), while only 47.1% of metaplastic TNBCs were positive for GATA3. When using MGP, GATA3, and TRPS1 as a novel IHC panel, 93.0% of breast carcinomas were positive for at least two markers, and only 9 cases were negative for all three markers. MGP was detected in 36 cases (3.0%) that were negative for both GATA3 and TRPS1. MGP showed mild-to-moderate positive expression in normal hepatocytes, renal tubules, as well as 31.1% (99/318) of hepatocellular carcinomas. Rare cases (0.6–5%) had focal MGP expression in renal, ovarian, lung, urothelial, and cholangiocarcinomas. Conclusions Our findings suggest that MGP is a newly identified sensitive IHC marker to support breast origin. MGP, TRPS1, and GATA3 could be applied as a reliable diagnostic panel to determine breast origin in clinical practice. Supplementary Information The online version contains supplementary material available at 10.1186/s13058-022-01569-1. | Matrix Gla protein (MGP), GATA3, and TRPS1: a novel diagnostic panel to determine breast origin
Metastatic breast carcinoma is commonly considered during differential diagnosis when metastatic disease is detected in females. In addition to the tumor morphology and documented clinical history, sensitive and specific immunohistochemical (IHC) markers such as GCDFP-15, mammaglobin, and GATA3 are helpful for determining breast origin. However, these markers are reported to show lower sensitivity in certain subtypes, such as triple-negative breast cancer (TNBC).
Using bioinformatics analyses, we identified a potential diagnostic panel to determine breast origin: matrix Gla protein (MGP), transcriptional repressor GATA binding 1 (TRPS1), and GATA-binding protein 3 (GATA3). We compared MGP, TRPS1, and GATA3 expression in different subtypes of breast carcinoma of (n = 1201) using IHC. As a newly identified marker, MGP expression was also evaluated in solid tumors (n = 2384) and normal tissues (n = 1351) from different organs.
MGP and TRPS1 had comparable positive expression in HER2-positive (91.2% vs. 92.0%, p = 0.79) and TNBC subtypes (87.3% vs. 91.2%, p = 0.18). GATA3 expression was lower than MGP (p < 0.001) or TRPS1 (p < 0.001), especially in HER2-positive (77.0%, p < 0.001) and TNBC (43.3%, p < 0.001) subtypes. TRPS1 had the highest positivity rate (97.9%) in metaplastic TNBCs, followed by MGP (88.6%), while only 47.1% of metaplastic TNBCs were positive for GATA3. When using MGP, GATA3, and TRPS1 as a novel IHC panel, 93.0% of breast carcinomas were positive for at least two markers, and only 9 cases were negative for all three markers. MGP was detected in 36 cases (3.0%) that were negative for both GATA3 and TRPS1. MGP showed mild-to-moderate positive expression in normal hepatocytes, renal tubules, as well as 31.1% (99/318) of hepatocellular carcinomas. Rare cases (0.6–5%) had focal MGP expression in renal, ovarian, lung, urothelial, and cholangiocarcinomas.
Our findings suggest that MGP is a newly identified sensitive IHC marker to support breast origin. MGP, TRPS1, and GATA3 could be applied as a reliable diagnostic panel to determine breast origin in clinical practice.
The online version contains supplementary material available at 10.1186/s13058-022-01569-1.
Breast cancer is the most common cancer among women worldwide. Approximately 5.8% of female breast cancers have distant metastasis at diagnosis (de novo stage IV breast cancer) [1]. The annual prevalence of recurrent metastatic breast cancer (initially diagnosed with stage I–III breast cancer and later progresses to stage IV breast cancer) is approximately three times that of de novo stage IV breast cancer [2]. Considering the high prevalence and high rate of metastasis of breast cancer, metastatic breast carcinoma is commonly considered during differential diagnosis when metastatic disease is detected in lymph nodes or organs such as the lung, liver, bone, and brain in females [3, 4]. In addition to the tumor morphology and documented clinical history, sensitive and specific immunohistochemical (IHC) markers are helpful to determine the breast origin. Currently, GATA-binding protein 3 (GATA3), gross cystic disease fluid protein 15 (GCDFP-15), and mammaglobin are commonly used breast cancer-specific IHC markers in clinical practice [5]. GATA3 is the most widely used breast-specific marker and has an overall sensitivity of > 90% [6–8]. GCDFP-15 and mammaglobin show lower sensitivities with high interstudy variation, which range from 40 to 75% and 40 to 70% [9–14], respectively. However, all these markers have been reported to show lower sensitivities in ER-negative breast cancer, especially in triple-negative breast cancer (TNBC), with sensitivities of GATA3 at 40–70% and GCDFP-15 and mammaglobin at < 30% [15–18]. Thus, there is a need to identify novel sensitive and specific markers and IHC panels for breast carcinomas. In the present study, we first identified six potential genes that were specifically upregulated in breast carcinoma, namely N-acetyltransferase 1 (NAT1), matrix Gla protein (MGP), secretoglobin family 2A member 2 (SCGB2A2/mammaglobin-A), LIM homeobox transcription factor 1 beta (LMX1B), transcription factor AP-2 beta (TFAP2B), and transcriptional repressor GATA binding 1 (TRPS1), by bioinformatics analyses using The Cancer Genome Atlas (TCGA) and Clinical Proteomic Tumor Analysis Consortium (CPTAC) databases across 24 different solid tumors. We further observed that only TRPS1 and matrix Gla protein (MGP) had comparable expression in luminal A/B, HER2-enriched, basal-like and normal-like subtypes of breast cancer. We next compared MGP, TRPS1, and GATA3 expression in 1201 breast carcinoma cases of different subtypes, including 140 metaplastic breast carcinoma cases, using immunochemistry staining. Moreover, as a newly identified marker to support breast origin, MGP expression was also evaluated in solid tumors (n = 2221) and normal tissues (n = 1351) from different organs.
TCGA mRNA expression data (transcripts per million [TPM] and raw count) were downloaded from the Gene Expression Omnibus (GEO) database under the accession number GSE62944 [19]. TCGA breast cancer PAM50 subtypes were obtained from Du et al. [20]. ER, PR, and HER2 status for TCGA breast cancer was obtained from Thennavan et al. [21]. Differential expression (DE) analysis was performed with the R package DESeq2 (version 1.30.0) [22], and genes with fold change > 4 and false discovery rate (FDR) < 0.05 were selected as upregulated DE genes. Overlapping upregulated DE genes between breast carcinoma (BRCA) and 23 other tumor types (ACC, BLCA, CESC, COAD, DLBC, GBM, HNSC, KICH, KIRC, KIRP, LAML, LGG, LIHC, LUAD, LUSC, OV, PRAD, READ, SKCM, STAD, THCA, UCEC, UCS, Additional file 2: Table S1) were defined as breast-specific genes. Genes with low expression (median TPM < 1) were further filtered.
mRNA expression data of 54 matched pairs of primary breast cancers and their metastases (brain, n = 22; bone, n = 11; ovary, n = 14; GI tract, n = 7) were downloaded from https://github.com/leeoesterreich/shiny-server/tree/master/apps/Paired_Mets [23–25]. The R package Limma (version 3.46.0) [26] was used to perform DE analysis with paired samples. Genes were defined as significantly downregulated in breast cancer metastasis compared to primary breast cancer if a fold change < 1 and FDR < 0.05 were found between all paired samples of primary breast cancers and breast metastases or between paired primary breast cancers and metastases from any specific site (brain, bone, ovary, or GI tract).
Normalized protein expression data of 105 TCGA breast cancer samples were downloaded from the CPTAC (https://cptac-data-portal.georgetown.edu/study-summary/S015) [27]. Spearman’s correlation between protein expression levels and mRNA expression levels (TPM) was performed using R v.4.0.3.
Tissue microarrays (TMAs) of breast carcinoma (invasive breast carcinoma of no special type, n = 1061), hepatocellular carcinoma (n = 318), ovarian carcinoma (n = 290), cholangiocarcinoma (n = 163), renal cell carcinoma (n = 121), lung adenocarcinoma (n = 297), colorectal adenocarcinoma (n = 346), gastric adenocarcinoma (n = 198), urothelial carcinoma (n = 218), thyroid carcinoma (n = 144), melanoma (n = 126), as well as normal tissues of liver (n = 159), ovary (n = 121), biliary duct (n = 159), kidney (n = 61), lung (n = 212), colorectum (n = 206), stomach (n = 137), bladder (n = 152), and thyroid (n = 144) were made in the Department of Pathology at the Sun Yat-sen University Cancer Center (SYSUCC). In addition, 140 formalin-fixed paraffin-embedded (FFPE) tissues from 129 patients with metaplastic breast carcinoma (MBC) were also included. All cases were previously diagnosed at SYSUCC during the year 2015–2020. For invasive breast carcinoma of no special type (NST) cases, the representative areas of both tumor and non-tumor normal tissues for each case were selected with 2:1 ratio and circled to match the blocks for the TMA analysis. For other tumor cases, two representative tumor areas for each case were selected for the TMA analysis. In addition, normal tissues from liver (n = 159), kidney (n = 61), ovary (n = 121), biliary duct (n = 159), lung (n = 212), colorectum (n = 206), stomach (n = 137), bladder (n = 152), and thyroid (n = 144) were also selected for the TMA analysis. This study was conducted in accordance with ethical standards and approved by the Ethics Committee of SYSUCC. Estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) status of breast carcinoma were routinely evaluated according to The American Society of Clinical Oncology/College of American Pathologists guideline recommendations [28, 29]. Herein, breast carcinomas were categorized into three groups based on ER, PR, and HER2 status as follows: ER and PR positive with HER2 negative (ER/PR+ group); HER2 positive regardless of ER and PR status (HER2+ group); and ER, PR, and HER2 negative (TNBC group).
Tissue microarrays and whole-tissue slides were stained with antibodies against MGP (mouse monoclonal, A-11; Santa Cruz Biotechnology, Dallas, TX, USA; dilution 1:100), TRPS1 (rabbit monoclonal, EPR16171; Abcam, Cambridge, MA, USA; dilution 1:800), and GATA3 (mouse monoclonal, EP368; Gene Tech, Shanghai, CHINA; dilution 1:600). Normal breast tissue samples were used as positive (incubated with primary antibody) and negative (incubated with antibody diluent) controls. MGP, TRPS1, and GATA3 IHC stains were reviewed and scored by pathologists PS, JH, XC, and ML. For MGP, cytoplasmic staining was considered positive. For TRPS1 and GATA3, only nuclear staining was considered positive. Immunoreactive scores were obtained by multiplying the percentage score (0, < 1%; 1, 1–10%; 2, 11–50%; 3, 51–100% cells positive) by the staining intensity (0, negative; 1, weak; 2, moderate; 3, strong) [30]. The degrees of MGP, TRPS1, and GATA3 expression were categorized as negative (0–1), mild positive (2), moderate positive (3–4), or high positive (6 and 9) based on immunoreactive scores (Fig. 1). Every case was reviewed and scored by two pathologists independently. Cases showing discrepant categorized results by the two pathologists were determined by discussion, and agreement was made by at least three pathologists.
Statistical analysis was performed using R (version 4.0.3). Categorical variables are presented as counts/total (percentage). The chi-square test of independence or two-sided Fisher exact test (if any cell of the 2 × 2 table had an expected count less than 5) was used to test whether two categorical variables (IHC positivity and gene) are related to each other. Post hoc Bonferroni correction was performed for multiple comparisons. A p value less than 0.05 was considered statistically significant.
A flowchart showing the identification of breast-specific candidate markers using a bioinformatic approach is shown in Fig. 2. We first performed differential gene expression analysis between breast carcinoma and 23 other tumor types (Additional file 2: Table S1). Thirty-three genes were commonly upregulated in breast carcinoma compared to other tumors, of which 19 genes had a median transcript per million (TPM) > 1. To identify genes highly expressed in both primary breast carcinoma and its metastasis, we removed genes that were significantly downregulated in metastatic breast carcinoma (Additional file 3: Table S2). Among the 12 remaining genes, six genes had a high correlation (correlation coefficient > 0.5, FDR < 0.05) between protein and mRNA expression levels (Additional file 4: Table S3), including LMX1B, TRPS1, NAT1, MGP, SCGB2A2, and TFAP2B, which could serve as potential IHC markers to support breast origin.
Since MGP and TRPS1 showed high expression in luminal A/B, HER2-enriched, TNBC/basal-like, and normal-like subtypes according to the PAM50 subtyping (Fig. 3, Additional file 1: Figure S1), we selected MGP, TRPS1, and GATA3 for further validation in 1201 cases of breast carcinoma of different subtypes and normal breast tissues. As shown in Fig. 1, GATA3 and TRPS1 showed nuclear staining while MGP showed cytoplasmic staining in the benign ductal epithelial cells, but not in myoepithelial cells. Immunoreactive scores of the enrolled cases are shown in Additional file 5: Table S4. MGP, TRPS1, and GATA3 expression in breast carcinomas is summarized in Table 1. MGP, TRPS1, and GATA3 were positive in 1075/1201 (89.5%), 1109/1201 (92.3%), and 922/1201 (76.8%) breast carcinomas, respectively. Comparable MGP and TRPS1 positivity was observed in HER2-positive (91.2% vs. 92.0%, p = 0.79) and TNBC subtypes (87.3% vs. 91.2%, p = 0.18), while MGP had relative lower positivity in ER/PR+ subtype (89.6% vs. 93.1%, p < 0.05). Although 93.4% of ER/PR+ tumors showed GATA3 expression, the positivity rates of GATA3 were lower than those of MGP or TRPS1 in HER2+ (77.0%, adjusted p < 0.001 for GATA3 vs. MGP and GATA3 vs TRPS1) and TNBC (43.3%, adjusted p < 0.001 for GATA3 vs. MGP and GATA3 vs TRPS1) subtypes. In ER/PR-positive tumors, most cases showed moderate–high expression of GATA3, which was more than those in TRPS1 or MGP (88.8% vs. 80.4% vs. 57.4%, adjusted p < 0.001 for GATA3 vs. MGP and GATA3 vs TRPS1). In contrast, the moderate–high positivity rate of GATA3 was significantly lower than that of TRPS1 or MGP in TNBC subtypes (30.3% vs. 80.6% vs. 68.0%, adjusted p < 0.001 for GATA3 vs. MGP and GATA3 vs TRPS1).
As shown in Table 1, MGP, TRPS1, and GATA3 were positive in 124/140 (88.6%), 137/140 (97.9%), and 66/140 (47.1%) TNBC-MBC patients, respectively. TNBC-MBC showed comparable positive staining of MGP (88.6% vs. 86.1%, p = 0.66) or GATA3 (47.1% vs. 39.6%, p = 0.24) than in TNBC of no special type (TNBC-NST) group. Notably, the positivity rate of TRPS1 was significantly higher in TNBC-MBCs than in TNBC-NSTs (97.9% vs. 84.7%, p < 0.001). According to histological subtypes, as shown in Table 2, MGP, GATA3, and TRPS1 were found to be positive in 86.5% (64/74), 62.2% (46/74), and 98.6% (73/74) of squamous cell carcinomas (SqCC), 83.3% (15/18), 22.2% (4/18), and 94.4% (17/18) of spindle cell carcinomas (SpCC), 92.5% (37/40), 32.5% (13/40), and 100% (40/40) of metaplastic breast carcinomas with mesenchymal differentiation (MBC-MD), and 100% (8/8), 37.5% (3/8), and 87.5% (7/8) of fibromatosis-like metaplastic carcinomas (FMC).
When combining GATA3, TRPS1, and MGP, we observed that 797/1201 (66.4%) of the enrolled breast carcinoma cases were positive for all three markers, including 452/565 (80.0%), 238/352 (67.6%), and 107/284 (37.7%) in the ER/PR+, HER2+, and TNBC subgroups, respectively (Table 3, Figs. 4 and 5). Among all enrolled cases, 36/1201 (3.0%; 7 ER/PR+, 13 HER2+, 16 TNBC) were positive for MGP, while TRPS1 and GATA3 were both negative; 40/1201 (3.3%; 10 ER/PR+, 7 HER2+, 23 TNBC) were positive for TRPS1, while MGP and GATA3 were both negative; 8/1201 (0.7%; 7 ER/PR+, 1 TNBC) were positive for GATA3, while MGP and TRPS1 were both negative. Only 9 cases (0.7%; 2 ER/PR+, 2 HER2+, 5 TNBC) were negative for all three markers. Different composite uses of MGP, TRPS1, and GATA3 may bring in GATA3-MGP, GATA3-TRPS1, MGP-TRPS1, and GATA3-MGP-TRPS1 IHC panels in daily practice. Assuming that the criteria used to determine the breast origin require at least two markers positive in these IHC panels, we found that the GATA3-MGP-TRPS1 panel showed the highest sensitivity (90.7%) among all enrolled cases, followed by MGP-TRPS1 (83.3%), GATA3-TRPS1 (72.9%), and GATA3-MGP (69.6%). In the TNBC and HER2+ groups, the sensitivity of the GATA3-MGP-TRPS1 (TNBC, 85.9%; HER2+, 93.2%) and MGP-TRPS1 (TNBC, 80.6%; HER2+, 84.1%) panels was significantly higher (adjusted p < 0.01 for all comparisons) than that of GATA3-MGP (TNBC, 38.7%; HER2+, 70.5%) and GATA3-TRPS1 (TNBC, 41.9%; HER2+, 73.9%). In ER/PR+ tumors, GATA3-MGP-TRPS1 demonstrated the best performance (96.5%, adjusted p < 0.001 for all comparisons), while the other three panels had comparable sensitivity at 84.1–87.8% (adjusted p > 0.05 for all comparisons); see Table 3 for more details.
Since the specificity of TRPS1 in breast carcinomas has been recently reported by Ai et al. [31], herein, as a newly identified marker to support breast origin, MGP expression was further evaluated in solid tumors (n = 2384) and normal tissues (n = 1351) from different organs (Table 4). MGP was negative in colorectal adenocarcinoma, gastric adenocarcinoma, thyroid carcinoma, and melanoma, while mild-to-moderate positivity was found in 31.1% (99/318) of hepatocellular carcinomas. Individual cases had focal MGP expression in renal cell carcinoma (6/121, 5.0%), ovarian carcinoma (7/290, 2.4%), lung adenocarcinoma (2/297, 0.7%), urothelial carcinoma (2/218, 0.9%) and cholangiocarcinoma (1/163, 0.6%), as shown in Fig. 6. In normal tissues, we observed that MGP showed mild-to-moderate positive expression in normal hepatocytes (159/159) and renal tubules (61/61) but the negative expression in other organs, including the ovary, biliary duct, lung, colorectum, stomach, bladder, and thyroid.
Specific IHC markers supporting the breast origin of an unknown carcinoma are important and helpful for diagnosis, especially ER-negative or triple-negative tumors. Currently, GATA3, GCDFP-15, and mammaglobin are commonly used panels to support breast origin, of which GATA3 is the most widely used. GATA-binding protein 3 (GATA3) is considered to be the most prevalent transcription factor involved in the proliferation and differentiation of ductal epithelium of the breast [32–34], which is linked to ER signaling [35, 36]. A number of studies have demonstrated that GATA3 is a superior marker for ER+ breast carcinoma than GCDFP-15 or mammaglobin, with a sensitivity consistently over 90% [7, 16, 35]. However, the sensitivity of GATA3 is significantly lower in ER-negative or TNBC subtypes, ranging from 15 to 60% in various studies [7, 16, 31, 35, 37, 38]. Our data also show that GATA3 exhibited extremely high sensitivity in 93.4% of ER/PR+ breast carcinomas, while up to 23.0% HER2 + BC and 60.4% TNBC were negative for GATA3. In addition, GATA3 is not a breast-specific marker that can label other common sources of tumors [39], including urothelial carcinomas, squamous cell carcinomas, lung adenocarcinoma, pancreatic adenocarcinomas, endocrine tumors, soft tissue sarcomas, and others. Currently, no single IHC marker is entirely breast-specific; GATA3 should be applied as part of an IHC panel, and more specific biomarkers are still required in the diagnostic setting. Using mRNA sequencing and proteomic data from TCGA and CPTAC of 24 different solid tumors, we identified six potential genes that are specifically upregulated in breast carcinoma: NAT1, MGP, SCGB2A2, LMX1B, TFAP2B, and TRPS1. Similar to the approach used by Ai et al. [29], the genes highly expressed in breast carcinoma compared to all other tumor types and equally highly expressed in all four PAM50 subtypes were identified as candidate markers. Moreover, our approach includes two more steps. Since gene expression may change during the process of metastasis, we also removed genes showing decreased expression in metastasis using RNA-Seq data from primary breast lesions and their paired metastases. In addition, we aimed to look for potential biomarkers that can be used in IHC-based assays, which are cellular protein labeling techniques; thus, we further selected genes with a high correlation between protein and RNA expression using protein expression data from CPTAC. SCGB2A2 (mammaglobin) and the most recently reported breast-specific marker TRPS1 are both listed in our final candidate markers, which suggests the robustness of our approach. This bioinformatic analysis process (Fig. 2) not only identifies MGP and TRPS1 as novel candidate IHC markers to support breast origin but also provides a new approach for the future selection of specific biomarkers in other tumor types. Trichorhinophalangeal syndrome type 1 (TRPS1) is named for a very rare hereditary disease with mainly autosomal dominant inheritance features characterized by craniofacial and skeletal abnormalities with damage and mutation affecting chromosome 8q [40]. TRPS1 is reported to be a transcriptional repressor that binds specifically to GATA sequences and represses the expression of GATA-regulated genes which function in vertebrate development, especially in the process of chondrocyte proliferation and differentiation [41, 42]. Some studies have suggested that TRPS1 may also act as a critical modulator in mammary epithelial cell growth, differentiation, and breast cancer development via epithelial–mesenchymal transformation, DNA replication, and mitosis [43, 44]. A recent study by Ai et al. reported that TRPS1 could serve as a sensitive and specific marker for breast carcinomas [31]. TRPS1 exhibited high sensitivity in ER/PR+ (98%), HER2+ (87%), and TNBC (86%) subtypes on TMAs. On the other hand, TRPS1 showed no or little expression in other tumor types. Parkinson et al. [37] and Yoon et al. [38] further verified the utility of TRPS1, showing higher sensitivity in the HER2+ and TNBC subgroups. Although the commercial TRPS1 antibody used in our study (EPR16171 from Abcam) is different from Ai’s and Parkinson’s study (TRPS: PA5-845874 from Invitrogen/Thermo Fisher), similarly high TRPS1 expression (91.2–93.1%) is also found in all types of breast carcinoma with the largest sample size thus far. Our findings confirm that both clones of TRPS1 are sensitive markers supporting breast origin. In addition, previously reported abnormal membranous expression of TRPS1 was not observed in our cohort. Previous studies reported that MGP is mainly secreted by chondrocytes [45, 46] and vascular smooth muscle cells [47], and it is considered a marker of vitamin K status in bone and vasculature, substantiating the role of MGP in extracellular matrix calcification regulation [48, 49]. MGP was recently found to be overexpressed in various types of cancer [50–52] and was reported to promote tumor progression by regulating angiogenesis [53]. In breast carcinoma, Yoshimura et al. and Gong et al. demonstrated that high MGP mRNA expression was associated with poor prognosis [52, 54]. However, whether MGP can serve as a breast-specific marker is unknown. In our cohort of 1201 breast carcinomas, every case matched benign breast ducts in a separate TMA or whole-slide section. MGP displays cytoplasmic labeling in nearly all ductal epithelial cells with various strengths but not in myoepithelial cells. Perivascular smooth muscle can be used as an internal positive control for MGP (Fig. 1). MGP was verified as a reliable marker with extremely high sensitivity in all subtypes of breast carcinoma (87.3–91.2%), which is comparable to TRPS1 and much higher than GATA3 in HER2+ and TNBC subtypes. We noticed that most MGP-positive cases demonstrated moderate and multifocal cytoplasmic staining patterns. There were generally more cases showing extensive and strong positivity for GATA3 and TRPS1 (greater than 49%) than for MGP (less than 40%, adjusted p < 0.001), except for the TNBC group (Table 1). Among the positive staining cases, 26.6% (286/1075, Table 1) showed mild positive of MGP, significantly higher than that of TRPS1 (14.1%, 157/1109, adjusted p < 0.001) and GATA3 (10.4%, 96/922, adjusted p < 0.001). More cases were categorized as mildly positive for MGP may be due to its cytoplasmic staining pattern, which is not preferable or easy to interpret subjectively like nuclear staining markers such as TRPS1 and GATA3. Mild cytoplasmic positivity tends to be more easily recognized as nonspecific or unstable as compared with mild nucleus positivity, which does affect the value of MGP as a single marker to determine breast origin in the clinical practice. Other commercial MGP antibodies could be also tested and verified in further studies. Even so, our data suggest that MGP has much better and more stable sensitivity than conventional nuclear (GATA3, SOX10 [38, 55, 56]) or cytoplasmic biomarkers (GCDFP15, mammaglobin) used to determine breast origin. The moderate–high positivity rate of MGP was significantly higher than that of GATA3 in TNBC-NST (65.3% vs. 34.8%, adjusted p < 0.001) and TNBC-MBC subtypes (70.7% vs. 17.4%, adjusted p < 0.001), suggesting the high sensitivity of MGP specially in the most troubling TNBCs. In addition, we observed that 239 GATA3-negative cases and 75 TRPS1-negative cases were positive for MGP, and 69 GATA3-mild positive cases and 96 TRPS1-mild positive cases showed moderate–high positive for MGP. Thus, using MGP, GATA3, and TRPS1 as a novel IHC panel significantly increased the sensitivity from 76.8–92.3% of the single marker (MGP, GATA3, or TRPS1) to 93.0–99.3% (≥ 1 positive or ≥ 2 markers positive for the GATA3, MGP, and TRPS1 panel). Although our IHC data of MGP were collected from primary breast carcinomas, our bioinformatics analysis revealed that MGP mRNA was not significantly changed between paired primary tumors and their metastases (Additional file 3: Table S2), which suggests that similar MGP expression could be found in metastatic breast carcinomas. Further verification of MGP is required in metastatic lesions as well as special types of invasive breast carcinoma, such as salivary gland-type tumors and neuroendocrine carcinoma. In the present study, we included 144 TNBC-NSTs and 140 TNBC-MBCs. GATA3 was expressed in only 39.6% of TNBC-NSTs and 47.1% of TNBC-MBCs and was mostly weakly positive, which is consistent with previous studies [31, 37, 38]. TRPS1 and MGP maintained high sensitivity in both TNBC-NSTs (84.7% and 86.1%) and TNBC-MBCs (97.9% and 88.6%). Focusing on MBCs, the sensitivity of TRPS1 (137/140, 97.9%) in our cohort was slightly higher than those reported by Ai et al. [31], Parkinson et al. [37], and Yoon et al. [38], which were 86.5% (45/52, adjusted p < 0.05), 91.0% (61/67, raw p = 0.061), and 95.0% (134/141, raw p = 0.33), respectively. TRPS1 exhibited a larger portion of strong positivity among the positive cases in TNBC-MBCs (112/137 [81.8%]) compared with the ER/PR+ (313/526 [59.5%], adjusted p < 0.001) or HER2+ (177/324 [54.6%], adjusted p < 0.001) group (Table 1). When MBCs were stratified by subtype, we observed that GATA3 showed relatively higher sensitivity in SqCCs (62.1%) than in other subtypes (22.5–37.5%). The majority of cases (85%, 34/40) in MBC with mesenchymal differentiation group showed chondroid/osseous differentiation, and both TRPS1 and MGP had the highest positivity in MBCs with chondroid/osseous differentiation (100% and 88.2%, respectively), followed by SqCCs and SpCCs. Interestingly, fibromatosis-like metaplastic carcinomas (FMCs) in our cohort were all positive for MGP (8/8). A total of 37.5% (3/8) and 87.5% (7/8) of FMCs were positive for TRPS1 and GATA3, respectively, which is inconsistent with Parkinson et al. [37] showing no single case with positive staining of TRPS1 or GATA3 in FMCs (0/3). In addition, we observed that 64 GATA3-negative MBCs and 2 TRPS1-negative MBCs were positive for MGP. The combined use of MGP, GATA3, and TRPS1 increased the sensitivity from 47.1–97.9% of the single marker (MGP, GATA3, or TRPS1) to 90.7–100.0% (≥ 1 positive or ≥ 2 markers positive for the GATA3, MGP, and TRPS1 panel) in MBCs. All these data suggest that both MGP and TRPS1 maintain excellent sensitivity in different subtypes of metaplastic breast carcinomas. However, TRPS1 may play a role in chondro-osseous differentiation. Wang et al. [57] found that TRPS1 was highly expressed in chondro-osseous sarcomas from both breast and extramammary sites, including heterologous components within malignant phyllodes tumors. Coincidentally, MGP is highly abundant in cartilage and acts as a critical regulator of calcification and turnover of bone and cartilage. The previous study showed that tumors exhibiting cartilaginous/osseous differentiation such as osteosarcoma and chondrosarcoma had high MGP expression [46, 58], and they can also metastasize to bone and lung-like breast cancer [59–62]. It would be hard to differentiate metaplastic breast carcinoma with cartilaginous/osseous differentiation from these tumors simply based on MGP positivity. Thus, pathologists should be cautious when faced with positive expression of MGP or TRPS1 in chondroid/osteoid components, especially with limited biopsy tissue. Further verification of MGP is also required in sarcomas and malignant phyllodes tumors. MGP was first isolated from bovine bone matrix in the 1980s [43]; since then, its expression has been demonstrated in normal endothelial cells, fibroblasts, chondrocytes, and vascular smooth muscle cells. Our data show that MGP had negative expression in normal organs, including the ovary, biliary duct, lung, colorectum, stomach, bladder, and thyroid. Interestingly, we observed that MGP was constantly expressed in normal hepatocytes, but the positive expression was detected in only 31.1% of hepatocellular carcinomas. In addition, previous studies also demonstrated that MGP was abundantly expressed in normal kidneys, specifically in the epithelium of Bowman's capsule and proximal tubules, where the activated protein contributes to maintaining renal microvascular traits [63, 64]. Consistently, we found that MGP was predominantly expressed in normal renal tubules, but the positivity rate significantly dropped to 5.0% in renal cell carcinomas. It is difficult for MGP itself to differentiate breast carcinoma from hepatocellular or renal cell carcinoma. The joint application of TRPS1 and GATA3 may be helpful since TRPS1 [31, 37] and GATA3 [7, 65] have been proven to be rarely positive in these tumors. Since high-grade ovarian serous carcinoma and breast carcinoma share similar morphologies and immunophenotypes, such as a micropapillary architecture and ER positivity, the diagnosis can be challenging. Our results showed that only 2.4% of ovarian serous carcinomas had focal MGP expression, which is lower than the reported positivity of GATA3 (~6%, [7, 66]) and TRPS1 (14%, [31]). Thus, this GATA3-MGP-TRPS1 panel may need inclusion with Pax-8 and WT-1 to differentiate breast carcinoma from serous carcinoma. Poorly differentiated lung adenocarcinomas have been frequently reported as TTF-1-negative and occasionally labeled for ER [67], while individual cases of breast carcinoma may show TTF-1 staining [68]. Thus, differentiating breast carcinoma and lung adenocarcinoma is common and sometimes difficult in clinical practice. In the current study, MGP was rarely expressed in lung adenocarcinomas (0.7%), which is lower than the previously reported positivity of GATA3 (∽8%, [7, 69]) and TRPS1 (2–3%, [31, 37]), indicating that MGP is a good marker to differentiate breast cancer from lung adenocarcinoma. We also found that MGP was positive in only 2 of 218 cases (0.9%) of urothelial carcinoma, which is known to be frequently labeled with GATA3 (70–90% [7, 70],). According to the documented literature and our data, the positivity of MGP, GATA3, and TRPS1 is extremely rare in other tumor types, such as cholangiocarcinoma and colorectal, gastric, and thyroid carcinomas. Ai et al. [31] found low TRPS1 expression in one melanoma, while none of the melanomas enrolled in our cohort was MGP-positive. Further investigation of MGP expression in other tumor types is needed, especially those for which relatively high TRPS1 or GATA3 expression has been reported, such as salivary duct carcinomas and pancreatic adenocarcinoma. A limitation of this study is that the TMA samples we used may not be able to adequately represent the intra-tumor expression heterogeneity of the IHC markers [62, 63]. A multicenter prospective study using standard whole-tissue sections should be undertaken to fully validate the value of MGP in determining breast origin. Our study used a relatively higher number of total breast carcinomas and metaplastic breast carcinomas than the recently published studies to identify new breast cancer marker [31, 37, 38]. Nevertheless, more cases are included in our ongoing study to further evaluate the sensitivity and specificity of MGP in metastatic breast carcinomas, special type of invasive breast carcinomas, neuroendocrine neoplasms, salivary gland-type tumors (either primary in breast or salivary gland), as well as tumors exhibiting cartilaginous/osseous differentiation.
Through bioinformatic analysis, we identified MGP as a novel IHC marker supporting breast origin, demonstrating relatively high sensitivity and specificity for invasive breast carcinoma of no special type. Further verification is needed for invasive breast carcinoma of special type as well as metaplastic breast carcinoma, especially those exhibiting cartilaginous/osseous differentiation. The joint application of MGP, TRPS1, and GATA3 could be recognized as a reliable diagnostic panel to determine breast origin in clinical practice.
Additional file 1: Fig. S1. The mRNA levels of candidate genes in different molecular subtypes of breast cancer. ER/PR+ (ER/PR+ and HER−, n = 685), HER2+ (n = 168), and TNBC (n = 177). TCGA BRCA ER, PR, and HER2 status were retrieved from Thennavan et al. [21].Additional file 2. Table S1. No. of cases in each TCGA tumor type.Additional file 3. Table S2. Expression change of 17 genes in paired breast cancers and their metastasesAdditional file 4. Table S3. Correlation between protein and mRNA levels of 9 candidate genes.Additional file 5. Table S4. The immunoreactive scores of GATA3, TRPS1 and MGP in 1201 cases of breast cancer. | true | true | true |
|
PMC9598232 | Jingfei Zhang,Qiming Li,Xin Zhang,Yanan Chen,Yufang Lu,Xinyu Wang,Lili Zhang,Tian Wang | Bisdemethoxycurcumin Alleviates Dextran Sodium Sulfate-Induced Colitis via Inhibiting NLRP3 Inflammasome Activation and Modulating the Gut Microbiota in Mice | 07-10-2022 | bisdemethoxycurcumin,colitis,inflammatory response,intestinal barrier function,apoptosis,NLRP3,gut microbiota | Our previous study showed that bisdemethoxycurcumin (BUR) exerts anti-inflammatory properties in lipopolysaccharide-induced intestinal injury, and studies have revealed that NOD-like receptor superfamily, pyrin domain containing 3 (NLRP3) inflammasome activation plays a vital role in the pathogenesis of colitis. However, it is not clear whether BUR could attenuate colitis-mediated intestinal inflammation via NLRP3 inflammasome inactivation and modulate the gut microbiota dysbiosis. The results demonstrated that BUR attenuated DSS-induced body weight decrease, histopathological changes, and epithelial apoptosis. BUR significantly improved the intestinal barrier defects and abrogated DSS-induced inflammatory response. Consistently, BUR reduced the expression of NLRP3 family members, confirming its inhibitory effects on NLRP3 inflammasome activation and pyroptosis. BUR regulated microbiota dysbiosis and altered the gut microbial community. BUR supplementation enriched the relative abundance of beneficial bacteria (such as Lactobacillus and Bifidobacterium), which showed significant negative correlations with the pro-inflammatory biomarkers. Collectively, these findings illustrated that BUR could ameliorate DSS-induced colitis by improving intestinal barrier function, reducing apoptosis, inhibiting NLRP3 inflammasome activation, and regulating the gut microbiota. | Bisdemethoxycurcumin Alleviates Dextran Sodium Sulfate-Induced Colitis via Inhibiting NLRP3 Inflammasome Activation and Modulating the Gut Microbiota in Mice
Our previous study showed that bisdemethoxycurcumin (BUR) exerts anti-inflammatory properties in lipopolysaccharide-induced intestinal injury, and studies have revealed that NOD-like receptor superfamily, pyrin domain containing 3 (NLRP3) inflammasome activation plays a vital role in the pathogenesis of colitis. However, it is not clear whether BUR could attenuate colitis-mediated intestinal inflammation via NLRP3 inflammasome inactivation and modulate the gut microbiota dysbiosis. The results demonstrated that BUR attenuated DSS-induced body weight decrease, histopathological changes, and epithelial apoptosis. BUR significantly improved the intestinal barrier defects and abrogated DSS-induced inflammatory response. Consistently, BUR reduced the expression of NLRP3 family members, confirming its inhibitory effects on NLRP3 inflammasome activation and pyroptosis. BUR regulated microbiota dysbiosis and altered the gut microbial community. BUR supplementation enriched the relative abundance of beneficial bacteria (such as Lactobacillus and Bifidobacterium), which showed significant negative correlations with the pro-inflammatory biomarkers. Collectively, these findings illustrated that BUR could ameliorate DSS-induced colitis by improving intestinal barrier function, reducing apoptosis, inhibiting NLRP3 inflammasome activation, and regulating the gut microbiota.
With the prevalence of irregular eating patterns and unfriendly environmental factors, inflammatory bowel diseases (IBD) have increased annually, including ulcerative colitis and Crohn’s disease, and have become one of the major threats to health globally [1]. The pathogenesis of IBD remains unclear, but recently researchers have confirmed that excessive oxidative stress plays an important role in the development of IBD [2]. Reactive oxygen species-related cell apoptosis and inflammasome activation aggravate intestinal barrier dysfunction and promote the progression of gut microbiota dysbiosis [3]. Dietary antioxidant intervention not only alleviates oxidant stress in inflamed intestine, but also regulates the immune−mediated inflammatory disorders [4,5,6]. In dextran sulfate sodium (DSS)−induced colitis, a well−accepted mouse model of IBD, natural antioxidants are reported to suppress inflammatory disorders through inflammasome inactivation and inhibition of inflammation. Therefore, supplementation with natural antioxidants may be a novel alternative treatment and intervention strategy for IBD. Inflammasome is an important component of innate immune response, and the NOD−like receptor superfamily, pyrin domain containing 3 (NLRP3) inflammasome is one of the most characterized inflammasomes. The NLRP3 inflammasome consists of a sensor molecule NLRP3, the adaptor apoptosis−associated speck−like protein containing a CARD (ASC), and the effector protease caspase 1 [7]. The NLRP3 inflammasome is activated by the oligomerization of NLRP3 with the ASC and pro−caspase 1 in a canonical pathway, which leads to the cleavage of caspase 1 and subsequently, the maturation of interleukin−1beta (IL−1β) and interleukin−18 (IL−18) [8]. These NLRP3−mediated inflammatory cascades are responsible for intestinal injury in response to inflammation [9]. Bauer and colleagues reported that NLRP3 contributed to inflammatory bowel disease in DSS−triggered colitis and was detrimental for intestinal epithelial barrier−maintenance [10]. Moreover, NLRP3 triggers pyroptosis via the N−terminal fragment of gasdermin−D (GSDMD), which forms pores in the plasma membrane, induces cell member rupture, and enhances an inflammatory response [11]. Recent studies showed that NLRP3 inactivation by siRNA or a specific inhibitor efficiently protected against intestinal barrier dysfunction and apoptosis, and alleviated colitis−induced cells’ pyroptosis [10,12]. Therefore, the inhibition of the NLRP3 inflammasome and NLRP3−initiated pyroptosis is a promising therapeutic strategy for the prevention and treatment of the inflamed intestine in colitis. Curcumin is a bioactive polyphenolic compound of turmeric (Curcuma longa), and widely used for traditional medicine in Indian and other Asian countries. Besides its excellent antioxidant activity, curcumin is reported to alleviate a wide spectrum of acute and chronic inflammatory disorders. Studies found that curcumin could suppress NLRP3 inflammasome activation in primary microglia and protect mice against ischemic stroke [13]. It was also indicated that curcumin alleviated DSS−induced colitis via inhibiting NLRP3 inflammasome activation and IL−1β production [14]. In doxorubicin−induced mice, curcumin inhibited NLRP3−mediated pyroptosis and enhanced the innate immune system [15]. Despite its excellent anti−inflammatory properties, there is limitation for the use of curcumin due to its poor absorption and bioavailability in vivo. Bisdemethoxycurcumin (BUR) is a derivative of curcumin and has more efficient pharmacological properties than curcumin [16]. In recent years, BUR has attracted increasing attention owing to its better bioavailability in vivo when compared with curcumin [17,18]. BUR is a minor constituent of curcuminoids and shows antioxidant, anti−inflammatory, antimutagenic, and antitumor effects [19]. Many studies have demonstrated that BUR has positive effects on organ protection, such as the liver, intestines, and kidneys. Our previous study reported that BUR was more potent than curcumin in alleviating circulating lipid peroxidation and facilitating the antioxidant gene expression of the liver and small intestine in broilers [20]. The latest research from our laboratory certified the protection of BUR against LPS-induced intestinal injury via reducing the inflammatory response. In a subsequent nutritional intervention study, we observed a decreased mRNA expression of intestinal IL-1β in broilers [21]. Moreover, it inhibited the neuroinflammation via the inhibition of the AKT/NF-κB and p38/JNK pathway in LPS-induced RAW 264.7 macrophage cells [22], and enhanced reno-protective effects and attenuated the toxicity of cisplatin in HK-2 cells and ICR mice [23]. However, whether BUR reduces IL-1β production via inhibiting NLRP3 inflammasome activation, and thereby attenuating inflammation has not yet been revealed. In this study, we established a DSS-induced intestinal inflammatory model in C57BL/6 mice to explore the effects of BUR in suppressing intestinal barrier dysfunction, moderating intestinal epithelial cell apoptosis, inhibition inflammatory response, and altering gut microbiota dysbiosis. Furthermore, the inhibitory effects of BUR on NLRP3 inflammasome activation and pyroptosis were further investigated.
All animal procedures were performed according to the guiding principles of the Animals Care and Ethics Committee of Nanjing Agricultural University, China (Certification No.: SYXK (Su) 2017–0007, 29 June 2021).
A total of 45 6−week−old male C57BL/6 mice (18–21 g) were purchased from Shanghai SLAC Laboratory Animal Ltd. Co. (Shanghai, China). The mice were caged in an environmentally controlled room (24 ± 1 °C, 12 h light–dark cycle and 50 ± 10% humidity) and had free access to water and food. Mice were randomly divided into four groups (six animals per group): the control group; DSS group; BUR (200 mg/kg) group; and BUR (400 mg/kg) group. The experimental timelines are shown in Figure 1A. Mice received corn oil, either 200 or 400 mg/kg body weight by gavage once per day for 14 days. On day 7–14, mice in the DSS and BUR groups were receiving DSS (2.5%) in drinking water to establish the colitis model. The dosage of DSS was chosen based on the previous reports [24,25]. The BUR used in the present study was provided by Kehu Bio−technology Research Center (purity ≥ 98%, Guangzhou, China). The doses of 200 and 400 mg/kg BUR used in the present study were according to our previous studies [21,24,26]. On day 15, all mice were sacrificed, and the serum and colons were collected. The colons were photographed, and the lengths of colons were immediately measured. Then, a small part of the colon was fixed in 4% paraformaldehyde for histological analysis; the remainder were snap frozen in liquid nitrogen and stored at −80 °C for subsequent analysis.
During the experimental period, the body weight, stool consistency, and fecal occult blood were recorded daily. The DAI score was calculated by the weight loss, stool consistency, and fecal occult blood as described previously (Table S1, Supplementary Materials) [27].
Interleukin−6 (IL−6), IL−1β, tumor necrosis factor−alpha (TNF−α), IL−18 in serum were measured using enzyme−linked immunosorbent assay kits (MultiSciences (Lianke) Biotech Co., Ltd., Hangzhou, China) according to the manufacturer’s instructions.
The fixed colon specimens were embedded in paraffin and cut into 5 μm slices. Then, the colon slices were stained with hematoxylin–eosin (H&E, Servicebio, China) and Alcian−blue/periodic acid Schiff reagent (AB−PAS, Servicebio, China) and observed using a light microscope.
The colonic MPO activity was measured using a commercial kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer’s instructions.
The apoptosis of the colon was detected by TUNEL assay using a TUNEL Bright Green apoptosis detection kit (Vazyme, Nanjing, Jiangsu, China). Briefly, the paraffin sections of the colon were dewaxed with xylene and rehydrated with ethanol. After treating with a proteinase K (20 μg/mL) for 20 min at room temperature, the tissue was incubated with 50 μL TdT buffer for 60 min at 37 °C in a dark and wet environment. Nuclei were then labeled with 4,6−di−amidino−2−phenylindole (Beyotime, Shanghai, China). Images were acquired using a confocal laser scanning microscope (Carl Zeiss) and analyzed by the Image−Pro Plus 6.0 software. The results were expressed as the ratio of the number of TUNEL−positive cells (green cell) to the total cell (blue cell).
Total RNA from colon samples were isolated using TRIzol Reagent (Takara, Dalian, China). The extracted RNA was calculated using a NanoDrop ND−1000 UV spectrophotometer (NanoDrop Technologies, Thermo Scientific, Waltham, MA, USA) at 260 and 280 nm and reverse−transcribed to cDNA using a PrimeScript RT Reagent kit (Takara Biotechnology Co., Dalian, China). The cDNA samples were amplified using the SYBR® Premix Ex Taq™ Kit (Takara Biotechnology Co. Ltd., Dalian, China) and a QuantStudio®5 real−time PCR Design & Analysis system (Applied Biosystems, USA) according to the manufacturer’s instructions. Briefly, the primer sequences are listed in Table S2. The relative expression levels were calculated with the 2 −∆∆Ct method. β−action was chosen as the internal control and used to normalize the results of target genes as described previously [24].
The colon tissues were homogenized with the ice−cold RIPA buffer containing protease inhibitors (Beyotime Biotechnology, Shanghai, China). The colon proteins were collected after centrifugation at 14,000× g for 5 min and determined using BCA protein assay kit (Beyotime Biotechnology, Shanghai, China). Equal amounts of protein were separated by 4–12% SDS−PAGE gel and transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA, USA). After blocking at room temperature with a commercial blocking buffer (Beyotime Biotechnology, Shanghai, China), membranes were incubated overnight at 4 °C with the following primary antibodies: Caspase 3, B cell lymphoma2 (Bcl2), Bcl2 associated X (Bax), cleaved Caspase 3 P17, cleaved Caspase 1 P20, and cleaved IL−1β, which were purchased from Affinity Biosciences (Cincinnati, OH, USA); Occludin, Claudin 1, ZO−1, ACS, NLPR3, Caspase 1, IL−1β, GSDMD, and β−actin, which were purchased from Proteintech (Wuhan, Hubei, China). Then, membranes were incubated with HRP−conjugated secondary antibody for 2–2.5 h at room temperature. Finally, the protein bands were visualized using an enhanced chemiluminescence (ECL) kit (Thermo Scientific, Wilmington, DE, USA) and detected by the ChemiDoc™ imaging system (BIO−RAD, Hercules, CA, USA). The results were quantified using Image J software and normalized to β−actin.
Total genomic DNA from twenty−four cecal digesta samples were extracted using the QIAamp Fast DNA Stool Mini Kit (Qiagen, MD, USA) according to the protocol. A PCR reaction was performed to amplify the V3–V4 region of the bacteria 16S rDNA using the specific primers with the barcode as previously described [21]. The PCR products were purified using the QIAquick Gel Extraction Kit (Qiagen, Germantown, MD, USA) and quantified using the QuantiFluor−ST Fluorometer (Promega, Madison, WI, USA). Then, sequencing was conducted with an Illumina Miseq PE300 platform (BIOZERON, Shanghai, China). According to the Mothur process (Ver 1.32.0, Ann arbor, MI, USA), data filtration was conducted to obtained qualified sequences. Sequences with a similarity of 97% were assigned to cluster the same operational taxonomic units (OTUs). Alpha diversity was analyzed to estimate the abundance and diversity of the bacterial community, including the Chao1, Shannon, and Simpson indices. Beta diversity analysis was performed by principal co−ordinate analysis (PCoA). Subsequent analysis of linear discriminant analysis effect size (LEfSe) was conducted to identify the significant and unique OTUs among groups.
The colonic digesta were homogenized with 1 mL distilled water and 0.2 μL metaphosphoric acid (25%, v/v). The mixture was stored at −20 ℃ for 30 min and then centrifuged for 10 min at 12,000 rpm. The supernatant was mixed vigorously with an equal volume of ice−cold ethyl ether and placed for 5 min at 4 ℃. The organic layers of samples were filtered with a 0.22 μm organic membrane and analyzed by gas chromatography as described previously [28,29].
The data were expressed as mean ± standard error mean (SEM). Comparisons among groups were conducted by one−way analysis of variance (ANOVA) followed by Tukey’s post−hoc test or the Kruskal–Wallis test with Dunn’s post−hoc test using the SPSS 22.0 software. Spearman’s rank correlation test was performed to evaluate the correlations between the relative abundance of top 10 bacteria at genus level and inflammatory−related mediators. A value of p< 0.05 was considered statistically significant.
To determine the effects of BUR on DSS−induced colitis, mice were treated by gavage with the control corn oil, 200 or 400 mg/kg BUR, respectively. During the 14−day experiment period, there was no mortality in any of the treatments. Weight loss after DSS treatment was significantly prevented by BUR (400 mg/kg) treatment in mice, but not by BUR (200 mg/kg) treatment (Figure 1B). BUR treatment significantly rescued the DSS−induced decrease in colon length (Figure 1C). Consistent with the data on body weight and colon length, BUR treatment significantly decreased the DAI scores (Figure 1D), suggesting the alleviated clinical colitis symptoms by BUR treatment. Additionally, serum IL−1β, TNF−α, IL−6, and IL−18 levels were analyzed. BUR pretreatment significantly reduced the levels of these inflammatory cytokines (Figure 1E–H).
Histological examination showed that the damaged crypts, increased infiltration of inflammatory cells, and goblet cells’ loss were observed in DSS group, were improved by BUR treatment (Figure 2A,B). BUR pretreatment significantly decreased the MPO activity of colon, confirming the anti−inflammatory effects of BUR in DSS−induced colitis. To further investigate whether BUR protected the intestinal barrier function, we measured the mRNA expression of tight junction proteins. The results showed that DSS suppressed the mRNA expression of occludin, claudin−1, and ZO−1, which were prevented by BUR treatment (Figure 2E–G). Furthermore, Western blotting demonstrated similar changes in protein levels, indicating that BUR could efficiently improve DSS−induced intestinal barrier integrity in mice (Figure 2H). These data suggested that BUR improved the intestinal barrier function via regulating the tight junction proteins.
Apoptosis is one of the mechanisms contributing to intestinal epithelial barrier dysfunctions and inflammatory response in DSS−induced colitis. We firstly determined whether BUR could attenuate apoptosis (Figure 3). According to the TUNEL assay, BUR pretreatment led to a significant decrease in the number of TUNEL−positive cells as compared to the DSS group (Figure 3A,B). We then confirmed whether the suppressive effect of BUR on apoptosis was mediated in a mitochondrial−dependent apoptotic pathway. The RT−PCR analyses showed that BUR pretreatment significantly suppressed the mRNA expression of Bax and caspase 3 while it promoted the mRNA expression of Bcl2 in DSS−treated mice. These observations were consistent with the Western blots’ results demonstrating that the protein expression of cleaved caspase−3 significantly reduced in both the 200 and 400 mg/kg BUR groups compared with the DSS group (Figure 3H). These results indicated that BUR potentially attenuated DSS−induced apoptosis via a Bax/Bcl2/caspase 3 pathway.
Considering that IL−1β and IL−18 were important NLRP3 inflammasome−dependent cytokines, and that the BUR treatment could reduce the release of these two inflammatory mediators, we then explored whether BUR inhibited NLRP3 inflammasome activation in colitis. The results showed that BUR treatment significantly decreased the mRNA and protein expression of NLRP3, ASC, and caspase−1 in DSS−treated mice (Figure 4A,B–E). Two doses of BUR produced a significant reduction in cleaved caspase 1 protein levels and blocked the maturation of IL−1β. Moreover, the mRNA and protein expression of GSDMD were obviously decreased with BUR (400 mg/kg) treatment. These results suggested that BUR could suppress NLRP3 inflammasome activation and induce pyroptosis through the NLRP3/Caspase−1/GSDMD signaling.
The bacterial communities were tested by Venn diagram analysis, and unique organisms are listed in Figure 5A. There were 217 unique organisms found in the control group, 82 in the DSS group, 156 in the BUR (200 mg/kg) group, and 23 in the BUR (400 mg/kg) group. The number of observed species and evaluation of the alpha diversity are shown in Figure 5C–F. However, there were no differences among the four groups in terms of α−diversity. Beta diversity analysis was presented by PCoA (Figure 5G). The PCoA of OTUs revealed that the colonic microbiota of the control group were distinctly separated from the DSS and BUR groups. The top 10 most abundant bacteria at the genus level in colon are presented in Table 1. Compared with the control group, BUR supplementation markedly increased the relative abundance of Dubosiella and Bifidobacterium (p < 0.05). We further performed the LefSe analysis to detect distinctive bacteria between the DSS and BUR groups (Figure S1 and Figure 6). The results (Figure S1) showed that Rikenellaceae, Eubacterium_siraeum_group, Rikenellaceae_RC9_gut_group, Rikenella, and Muribaculaceae were differentially enriched in the BUR (200 mg/kg) group, whereas Allorhizobium_Neorhizobium_Pararhizobium_Rhizobium, Lachnospiraceae_FCS020_group, Hydrotalea, Sphingomonadaceae, Sphingomonadales, Sphingomonas, Asinibacterium, Streptococcus, and Streptocaccaceae were differentially enriched in the DSS groups. Compared with the DSS group, the BUR (400 mg/kg) group had higher relative abundances of Dubosiella, Actinobacteriota, Actinobacteria, Bifidobacterium, Bififobacteriales, Bififobacteriaceae, Atopobiaceae, Olsenella, Paraburkholderia, Burkholderiaceae, Tissierellales, Eubacterium_nodatum, Parvibacter, Coriobacteriaceae_UCG_002, Peptostreptococcaceae, and Romboutsia. The concentrations of total acids including acetate, propionate, and butyrate showed a notable reduction in DSS−treated mice (Figure 7). BUR treatment significantly improved the colonic SCFAs’ profiles. BUR (400 mg/kg) treatment significantly increased the concentrations of intestinal SCFAs but only the concentrations of propionate and butyrate were remarkably improved at 200 mg/kg BUR.
To further determine the potential relationship between gut microbiota and the inflammatory response, we performed Spearman’s rank correlation analysis between the top 10 most abundant bacteria at the genus level and the inflammatory mediators (Figure 8). The results showed that Muribaculaceae_norank showed significant positive correlations with serum IL−18, MPO activity, and NLRP3 protein expression. Lactobacillus had significant positive correlations with the acetate and total acid concentrations while it had negative correlations with serum IL−6 and IL−18, MPO activity, and ASC protein expression. The abundance of Burkholderia−Caballeronia−Paraburkholderia had significant positive correlations with serum IL−6 and IL−1β. Desulfovibrio was positively correlated with serum TNF−α and ASC protein expression.
The bio−function of gut microbial communities corresponds to their compositional changes. Thus, we performed the functional analysis of intestinal microbial community by PICRUST. As shown in Figure 9A, 26 KEGG pathways of DSS−treated mice were found to be changed as compared to those of mice in the control group. After BUR treatment, three KEGG were improved compared with the DSS group including flavone and flavonol biosynthesis, and fatty acid biosynthesis in the BUR groups (Figure 9B,C).
With the increasing incidence and prevalence globally, IBD has been perceived as one of the major problems threatening human health. Multiple studies have highlighted the pathogenic role of immune−mediated inflammatory disorders in IBD through mechanisms involving intestinal barrier defects, inflammasome activation, and gut microbiota dysbiosis [30]. In this study, we found that BUR supplementation alleviated DSS−induced colitis in mice by improving intestinal barrier function, inhibiting apoptosis, blocking the inflammatory signal, and modulating the gut microbial community. Importantly, the protective effects of BUR were involved in suppressing the maturation of pro−inflammatory cytokines and inhibiting the pyroptosis via NLRP3 inflammasome inactivation. The classical DSS−induced colitis in rodent models with morphological changes and inflammatory disorder is consistent with the clinical symptoms of IBD, including body weight loss, rectal bleeding, shortened colon length, and elevated DAI scores [31]. In the current study, we found the above−mentioned manifestations were effectively alleviated by the 200 and/or 400 mg/kg BUR supplementation. BUR pretreatment prevented the DSS−induced body weight loss, colon length reduction, and DAI score increase. In addition, the excessive release of pro−inflammatory cytokines into the systemic circulation characterizes a systemic inflammatory reaction as colitis occurs, which provides reliable indicators for the disease severity. BUR supplementation significantly reduced serum IL−6, TNF−α, IL−1β, and IL−18 to approximately the control levels. These results confirmed the protective effects of BUR in DSS−induced colitis. In rodents with colitis, the abnormal inflammatory cells’ infiltration is enhanced because of the intestinal barrier defects, and the goblet cells are also lost due to the disruption of the intestinal epithelial layer [32]. Our histological results showed that BUR treatment not only alleviated histopathologic damage induced by DSS, but also mitigated the goblet cells’ loss. In parallel, the reduced MPO activities in the colon were observed following BUR pretreatment, indicating the beneficial effects of BUR in the remission of colitis. The intestinal barrier represents an important physical mechanism that provides a defense against pathogens and endotoxins, which, in turn, can be linked with intestinal epithelial integrity and immune−mediated inflammatory responses [33]. In the present study, the decreases in tight junction (TJ) molecules, including occludin, claudin−1, and ZO−1, led to intestinal barrier dysfunction and promoted a strong intestinal inflammation. Curcumin and BUR were demonstrated to prevent intestinal tight junction dysfunction caused by leptin exposure, hypobaric hypoxia injury, and cisplatin poisoning [34,35,36,37]. Our results were consistent with previous studies demonstrating that BUR treatment inhibited the decreased mRNA expression of ZO−1, occludin, and claudin−1 genes caused by DSS. In BUR groups, the protein expression of ZO−1, occludin, and claudin−1 were restored as well. These results suggested that BUR treatment reduced the intestinal barrier impairment and relieved the colonic damage via regulating the TJ molecules. One of the remarkable findings of this study was that we determined the induction of apoptosis in colitis and BUR had an obvious protection against apoptosis. Apoptosis is a form of programmed cell death that implicated in the pathogenesis of colitis [38]. Massive intestinal epithelial apoptosis leads to the dysregulation of the intestinal barrier integrity, and thus triggers a severe inflammatory response in colitis. Our TUNEL staining demonstrated the extensive apoptosis in the DSS−treated colon. BUR supplementation decreased the mRNA expression of pro−apoptotic Bax and increased that of anti−apoptotic Bcl2. A few studies have revealed the effects of BUR on the apoptosis. Research has clarified that BUR abolished cisplatin−induced apoptosis in renal tubular epithelial cells and suppressed apoptosis induced by staurosporine and tert−butyl hydroperoxide [18,23,39]. The Bcl2 family proteins participate in the classical mitochondrial−dependent apoptotic pathway that mediates the cell death [40]. The induction of Bax leads to the release of cytochrome c and subsequently activates the downstream caspase 3, which degrades protein substrates to form the terminal cleavage and contributes to the DNA fragmentation [41]. By Western blotting analysis, we confirmed the inhibitory effects of BUR on the activation of caspase 3. Because of the reduction in apoptosis−executing molecules such as caspase 3 and cleaved caspase 3, BUR served as an effective apoptotic inhibitor against colitis, which might explain the mechanism by which BUR maintained the intestinal barrier integrity and blunted the DSS−induced inflammation. Previous studies showed that the induction of apoptosis was required for the activation of the NLRP3 inflammasome in the signaling cascade, but the concrete contribution of NLRP3 inflammasome activation to DSS−induced colitis was still elusive. There were some reports about the potential involvement of the NLRP3 inflammasome in intestinal epithelial cells after DSS−induced colitis [10,42]. In this study, BUR was found to inhibit NLRP3 inflammasome activation in DSS−treated mice. Consistent with this notion, Gong (2018) et al. found that curcumin strongly suppresses DSS−induced NLRP3 inflammasome activation and alleviates colitis in mice [14]. The less−cleaved caspase−1 and ASC specks by pretreatment produced the inhibitory effects of curcumin on the NLRP3 inflammasome activation and IL−1β secretion in LPS−primed peritoneal macrophages and BMDMs [14]. The activation of the NLRP3 inflammasome depends on two signals: priming and activation. The priming signal is achieved by the NLPR3 and IL−1β over−expression mediated by external stimuli [7]. First, our data revealed a decreased expression of NLRP3 and IL−1β in BUR groups compared with the DSS group, suggesting that BUR diminished the first signal required by the NLRP3 inflammasome activation. The activating signal requires the assembly of NLRP3 inflammasome components comprising NLRP3, ASC, and pro−caspase1, which promote caspase−1−mediated IL−1β and IL−18 secretion and pyroptosis [43]. Thus, the involvement of caspase−1 and GSDMD, a biomarker of pyroptosis, was also determined to investigate whether BUR inhibited the second assembly signal of the NLRP3 inflammasome. We found that BUR decreased the protein expression of caspase 1 and blocked the activation of caspase 1. The formation of the NLRP3 inflammasome complex requires activated caspase−1 involvement. The activation of caspase−1 cleaves GSDMD to trigger pyroptosis mediated by the mature NLRP3 inflammasome [44]. Pyroptosis leads to cell death coupled to significant increases in IL−1β and IL−18, which in turn promote the inflammatory process [45]. The results demonstrated that BUR pretreatment decreased the GSDMD protein levels, which provided another explanation for the anti−inflammatory activity of BUR related to its inhibitory role on the NLRP3 inflammasome pathway. In addition, BUR−induced apoptosis inhibition played an indirect role in the induction of NLRP3 inflammasome activation due to the diminished release of IL−1β mediated by the reduction in Bcl2 [46]. The DNA fragment, especially the damaged mitochondrial DNA, is an activator of the NLRP3 inflammasome activation [46]. We assumed that suppression of Bcl2 expression and the release of DNA fragments mediated, at least in part, the inhibitory effects of BUR on the NLRP3 inflammasome activation. The mechanisms underlying the beneficial effects of BUR were involved in both the inhibition of apoptosis and NLRP3 inflammasome activation, thus resulting in improved intestinal barrier function and reduced inflammatory injury. The gut microbiota plays an important role in maintaining intestinal barrier function and is a sensitive contributor to the epithelium cell apoptosis and inflammation. The gut microbiota dysbiosis is responsible for the pathogenesis of IBD, and mediates the development of inflammatory reaction. The DSS challenge seriously disrupts the balance of the colon microbiome. BUR did not affect the alpha or beta diversity of the bacterial communities reduced by DSS. However, at the genus level, the abundances of multiple bacteria were rebalanced in the BUR−treated groups. Lactobacillus is a beneficial bacteria prevalent in the healthy intestine and depleted in IBD patients. Enriched Lactobacillus has been reported to improve intestinal barrier integrity and restrain pathogen growth, resulting in the increase in SCFAs producers and an anti−inflammatory response. In agreement, our HPLC analysis of colonic digesta showed that the DSS challenge significantly decreased the production of gut microbiota−produced SCFAs. The concentrations of acetate, propionate, and butyrate in the BUR groups were higher than those of the control group. There was a mutual regulation between SCFAs and gut inflammation. The enhanced SCFAs production induced by BUR supplementation might prevent the inflammation caused by DSS. Moreover, the LEfSe analysis showed that the BUR groups had higher populations of Bifidobacterium compared with the DSS group. Bifidobacterium assists in promoting gut health, and commonly serves as a potential probiotic and an alternative intervention of animal nutrition. Previous studies found that high digestive abundance of Bifidobacterium and Lactobacillus contributed to improving intestinal inflammation by suppressing NF−kappaB, and they could restrict the colonization of pathogenic bacteria by lowering gastrointestinal tract pH. The increased abundances of health−promoting taxa in the BUR groups showed the potential application of BUR in treating colitis by regulating the gut microbiota. Growing evidence has shown that the alteration of bacterial communities was closely associated with gut disease severity, especially the intestinal inflammatory status. Analogously, in the present study, the correlation analysis between the top 10 abundant genus bacteria and a wide spectrum of inflammation−related biomarkers was determined. The results showed that the abundance of Lactobacillus had a highly positive correlation with the SCFAs profiles while it had negative correlations with inflammatory factors, which were conducive to the suppression of intestinal inflammation. Lachnospiraceae_NK4A136_group is a kind of SCFAs−producing bacteria that promotes the biosynthesis of butyric acid [47,48]. We found that the abundance of the Lachnospiraceae_NK4A136_group was significantly increased in the DSS group, whereas its increase was also observed in the BUR groups. It was assumed that the improved SCFAs profiles were related to the enrichment of the Lachnospiraceae_NK4A136_group. Dubosiella may be a potentially beneficial bacteria and its enrichment helps protect in some dietary treatments against colitis [49]. Although DSS induced the increase in Dubosiella, we also demonstrated a similar increased abundance of Dubosiella in BUR groups, consistent with previous reports [50,51]. The enrichment of Burkholderia−Caballeronia−Paraburkholderia and Desulfovibrio were positively associated with the biomarkers that related to the NLRP3 inflammasome pathway. The role of Burkholderia−Caballeronia−Paraburkholderia in colitis is controversial. Wang (2022) et al. found that Burkholderia−Caballeronia−Paraburkholderia belongs to the proteobacteria phylum and is the dominant bacteria in the model mice of DSS−induced colitis, which was similar to our results [52]. Chibuzor−Onyema (2021) et al. reported that Burkholderia−Caballeronia−Paraburkholderia exhibits similar lactic acid−producing effects as Lactobacillus and Lactococcus [53]. An increased abundance of Burkholderia−Caballeronia−Paraburkholderia were also observed in the BUR groups, indicating that these conflicting results warranted further exploration. Desulfovibrio is a sulfate−reducing bacteria and frequently causes human infection by generating LPS [54,55]. As reported, the abundance of Desulfovibrio was decreased in resveratrol−treated high−fat diet mice, which was consistent with our results showing an improved gut barrier function in BUR groups [56]. The severity of the intestinal inflammatory response was decreased after BUR treatment in colitis, and the abundance of harmful bacteria (Muribaculaceae_norank and Desulfovibrio) were also decreased. BUR had a good anti−inflammatory activity and contributed to maintaining the abundances of Bifidobacterium and Limosilactobacillus. We speculated that the changed abundance of gut microbiota following BUR supplementation might avail in the regulation of inflammatory status and contribute to the remission of colitis. Of note, the predicted function of the changed gut microbiota were obtained with a phylogenetic reconstruction of unobserved states (PICRUSt) analysis. The function of the “two−component system” and “ABC transporters” were enriched in the DSS group compared with the control group. The ABC transporters are involved in the uptake of micronutrients and the export of cytotoxic compounds for gut microbiota [57]. BUR (200 mg/kg) supplementation enhanced the function of “fatty acid biosynthesis” and BUR (400 mg/kg) enhanced the function of “flavone and flavonol biosynthesis” in DSS−challenged mice. Based on these conflicting results of two doses, we speculated that the effects of BUR on gut microbiota were not in a concentration−dependent manner. To be honest, the results of gut microbiota are rarely consistent since distinct pathways are implicated in the regulation of natural extracts including BUR. BUR is a natural polyphenol derived from the herbal root of Curcuma longa. Once administrated, BUR metabolism mainly occurs via reduction and conjugation that generate various degradation and metabolism products. The assumption that natural compounds such as BUR exert a broad spectrum of biological and cellular activities directly, or indirectly through their primary and secondary metabolites, may provide an explanation for the complicated mechanism of BUR in vivo. BUR concentrations are rarely detectable in plasma at 24 h with low dietary inclusion levels, suggesting its rapid systemic metabolization and elimination. We consider that the indirect effect derived from various metabolites play a primary role in the regulation of the signaling pathway, such as fatty acid biosynthesis in the BUR (200 mg/kg) group. Increased dietary inclusion level of BUR may be related to the induction of flavone and flavonol biosynthesis because elevated BUR accumulation in serum and digesta has been observed in the BUR (400 mg/kg) group.
In conclusion, our study demonstrated that BUR alleviated DSS−induced colitis by improving intestinal barrier function, reducing apoptosis, and decreasing inflammation through the NLRP3 inflammasome inactivation. In addition, BUR treatment improved the gut microbial community towards exerting anti−inflammatory effects and promoting host health. These results indicated that BUR might represent a promising protective agent against intestinal damage and gut microbiome dysbiosis for the treatment and therapy of IBD. | true | true | true |
|
PMC9598437 | Hui-Jeong An,Sung-Hwan Cho,Han-Sung Park,Ji-Hyang Kim,Young-Ran Kim,Woo-Sik Lee,Jung-Ryeol Lee,Seong-Soo Joo,Eun-Hee Ahn,Nam-Keun Kim | Genetic Variations miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G and the Risk of Recurrent Pregnancy Loss in Korean Women | 25-09-2022 | recurrent pregnancy loss,single-nucleotide polymorphism (SNP),microRNA | This study investigated the genetic association between recurrent pregnancy loss (RPL) and microRNA (miRNA) polymorphisms in miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G in Korean women. Blood samples were collected from 381 RPL patients and 281 control participants, and genotyping of miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G was carried out by TaqMan miRNA RT-Real Time polymerase chain reaction (PCR). Four polymorphisms were identified, including miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G. MiR-10a dominant model (AA vs. AT + TT) and miR-499bGG genotypes were associated with increased RPL risk (adjusted odds ratio [AOR] = 1.520, 95% confidence interval [CI] = 1.038–2.227, p = 0.032; AOR = 2.956, 95% CI = 1.168–7.482, p = 0.022, respectively). Additionally, both miR-499 dominant (AA vs. AG + GG) and recessive (AA + AG vs. GG) models were significantly associated with increased RPL risk (AOR = 1.465, 95% CI = 1.062–2.020, p = 0.020; AOR = 2.677, 95% CI = 1.066–6.725, p = 0.036, respectively). We further propose that miR-10aA>T, miR-30cA>G, and miR-499bA>G polymorphisms effects could contribute to RPL and should be considered during RPL patient evaluation. | Genetic Variations miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G and the Risk of Recurrent Pregnancy Loss in Korean Women
This study investigated the genetic association between recurrent pregnancy loss (RPL) and microRNA (miRNA) polymorphisms in miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G in Korean women. Blood samples were collected from 381 RPL patients and 281 control participants, and genotyping of miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G was carried out by TaqMan miRNA RT-Real Time polymerase chain reaction (PCR). Four polymorphisms were identified, including miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G. MiR-10a dominant model (AA vs. AT + TT) and miR-499bGG genotypes were associated with increased RPL risk (adjusted odds ratio [AOR] = 1.520, 95% confidence interval [CI] = 1.038–2.227, p = 0.032; AOR = 2.956, 95% CI = 1.168–7.482, p = 0.022, respectively). Additionally, both miR-499 dominant (AA vs. AG + GG) and recessive (AA + AG vs. GG) models were significantly associated with increased RPL risk (AOR = 1.465, 95% CI = 1.062–2.020, p = 0.020; AOR = 2.677, 95% CI = 1.066–6.725, p = 0.036, respectively). We further propose that miR-10aA>T, miR-30cA>G, and miR-499bA>G polymorphisms effects could contribute to RPL and should be considered during RPL patient evaluation.
Recurrent pregnancy loss (RPL) is generally defined as three or more consecutive losses of pregnancy before 20 weeks of gestation. However, the American Society for Reproductive Medicine recently redefined RPL as more than two consecutive pregnancy losses [1]. Worldwide, RPL is a serious health problem that is significantly associated with morbidity and mortality. Factors contributing to the etiology of RPL include advanced maternal age, maternal anatomic anomalies, placental anomalies, chromosomal abnormalities, endocrine dysfunction, antiphospholipid syndrome, hereditary thrombophilia, psychological trauma, and environmental factors, such as smoking, excessive alcohol consumption, and stress [2]. Additionally, women who miscarry during their first pregnancy are 5% more likely to develop RPL than healthy women [3]. Although many relevant factors have been identified, the root cause of most cases of RPL remains unknown. RPL is also associated with blood clotting angiogenesis and immune disorders. MicroRNAs (miRNAs) are small (approximately 23 nucleotides), noncoding, single-stranded RNA molecules that form base pairs with complementary target messenger RNAs (mRNAs) [4]. It has been demonstrated that miRNAs modulate gene expression via destabilization or translational repression of target mRNAs [5,6]. Furthermore, miRNAs have been implicated in the regulation of several biochemical pathways in various eukaryotic organisms [7,8]. RNA polymerase II transcribes miRNAs into long precursor transcripts known as primary (pri)-miRNAs, which are subsequently converted into pre-miRNAs by DROSHA, which is a ribonuclease type III enzyme that forms a functional complex with DiGeorge syndrome critical region 8 [9,10]. The pre-miRNA is then exported to the cytoplasm by the exportin5 (XPO5)-RAS–related nuclear protein (RAN)-guanosine-5′-triphosphate (GTP) complex [11]. RAN is a small GTP-binding protein, and the RAN GTPase-XPO5 complex forms a heterotrimer with the pre-miRNA [12]. The pre-miRNA is processed by RNase III DICER to release the miRNA duplex, which is a double-stranded RNA approximately 23 nucleotides in length. DICER also initiates the formation of the RNA-induced silencing complex (RISC) [13], which is responsible for miRNA-mediated gene silencing and RNA interference. The biological function of the miRNA is initiated by binding to the 3′-untranslated region (UTR) of the target mRNA, thereby repressing its expression. A single miRNA can regulate the expression of multiple target mRNAs, thus serving as a master controller of gene expression. Multiple studies have recently demonstrated the roles of miRNAs in the pathophysiology of several ovarian diseases, including polycystic ovary syndrome (PCOS) and primary ovarian insufficiency (POI) [14,15]. POI, which is also known as premature ovarian failure, is characterized by insufficient or premature depletion of ovarian reserves, which leads to infertility [16]. The findings of the present study suggest that miRNAs play an essential role in the normal function and regulation of reproductive organs. The expression of a given gene may be affected or regulated by its genetic variations, and single-nucleotide polymorphisms (SNPs) are the most common genetic variation affecting DNA [17]. SNPs or mutations in genes encoding miRNAs can affect miRNA properties, resulting in their altered expression and/or maturation [18]. Sequence variations around the processing sites of miRNAs or in the mature miRNA itself, particularly in the seed sequence, can profoundly affect miRNA biogenesis and function [19]. Polymorphisms in pre-miRNAs were first reported in 2005 [20], and several studies on the associations of these polymorphisms have since been reported [21,22]. Aberrant miRNA expression has been implicated in numerous diseases; therefore, considerable research efforts are currently being made for miRNA-based therapies [23]. In the present study, we performed a database search and identified four SNPs in pre-form miRNAs: miR-10aA>T (rs3809783), miR-30cA>G (rs113749278), miR-181aT>C (rs16927589), and miR-499bA>G (rs3746444). All of these miRNAs are reportedly associated with various reproductive diseases [24,25,26,27]. Therefore, we hypothesized that the SNPs miR-10a, miR-30c, miR-181a, and miR-499b play a role in the development of RPL. The minor allele frequency of these SNPs is >5% in the Asian population; however, whether they are genetically associated with RPL or whether miRNA expression varies as a function of these pre-form polymorphisms remains unclear. We, therefore, investigated the correlation between RPL and these miRNA polymorphisms.
Blood samples were collected from 381 RPL patients (mean age ± standard deviation [SD], 33.00 ± 5.73 years) and 281 control participants (33.03 ± 4.36 years). Blood samples were collected prior to 20 weeks of gestation based on human chorionic gonadotropin (hCG) levels. The RPL patients were recruited from the Department of Obstetrics and Gynecology or the Fertility Center at the CHA Bundang Medical Center in Seongnam, South Korea between March 1999 and February 2010. Women in the control group were recruited from CHA Bundang Hospital and met the following criteria: history of at least one spontaneous pregnancy; current pregnancy; regular menstrual cycles; karyotype 46, XX; and no history of miscarriage. The study abided by the Declaration of Helsinki and was approved by the Institutional Review Board of CHA Bundang Medical Center (IRB approval no. BD2010-123D), and written informed consent was obtained from all participants. All RPL patients had suffered a minimum of two consecutive spontaneous miscarriages at an average gestational stage of 7.36 ± 1.93 weeks. Pregnancy loss was diagnosed based on the results of hCG tests, ultrasound, and/or physical examination before 20 weeks of gestation. None of the participants had a history of smoking or alcohol use. The following parameters were also measured: activated partial thromboplastin time (aPTT), body mass index (BMI), blood urea nitrogen (BUN), creatinine, estradiol (E2), follicle-stimulating hormone (FSH), luteinizing hormone (LH), platelet (PLT) count, and prothrombin time (PT), using participant blood samples. Patients with the following conditions were excluded from the study: RPL or implantation failure due to hormonal, genetic, anatomic, infectious, autoimmune, or thrombotic causes. Anatomic causes were evaluated using hysterosalpingogram, hysteroscopy, computed tomography, and magnetic resonance imaging to detect intrauterine adhesions, septate uterus, and uterine fibroids. Hormonal causes, including hyperprolactinemia, luteal insufficiency, and thyroid disease, were evaluated by blood analyses. Infectious causes, such as the presence of Ureaplasma urealyticum or Mycoplasma hominis, were evaluated by bacterial culture. Autoimmune causes, including antiphospholipid syndrome or lupus, were evaluated using lupus anticoagulant and anticardiolipin antibodies. Thrombotic causes, such as thrombophilia, were evaluated by identification of protein C and S deficiencies and by detection of β-2-glycoprotein 1 antibodies.
A total of 150 μL of whole blood and fluorochrome-labeled monoclonal antibodies against anti-CD3-FITC (1:100, 555339), anti-CD4-PE(1:100, 357404) anti-CD8-PE-cy5 (1:20, 344769) anti-CD19-APC (1:100, 392503), anti-CD56-PE-Cy7 (1;100, 392411) NK cells were added to each tube. All antibodies were obtained from Biolegend (San Diego, CA, USA). The tubes were vortexed and incubated in the dark at room temperature for 40 min. Next, 2 mL of Lyse solution (diluted 1:10; BD Bioscience, Sunnyvale, CA, USA) was added, and the tubes were vortexed again, incubated at room temperature for 30 min, and centrifuged at 1200 rpm for 5 min. The cells were then washed three times with 2 mL of PBS each wash, and the cells were suspended in 250 μL of PBS and analyzed by flow cytometry (BD Bioscience).
Chromosome analysis was conducted according to standard cytogenetic methods. Peripheral blood lymphocytes were cultured for 70 h, and then KaryoMAX Colcemid Solution (Gibco) was added when the chromosomes were at the metaphase stage. KCl (0.05 M) was added as a hypotonic agent, and the cells were fixed for harvest using a fixative formed by adding one volume of acetic acid to two volumes of methanol. Metaphase chromosome preparations obtained after cell culture were stained using the Giemsa-Trypsin-Giemsa (GTG) banding method.
Genomic DNA was extracted from anticoagulant-treated peripheral blood samples using a G-DEX Genomic DNA extraction kit (iNtRON Biotechnology, Seongnam, Korea) [28,29]. Briefly, Proteinase K was added to a microcentrifuge tube, followed by 30 µL of blood. Next, 300 µL of Lysis solution was added, and the samples were vortexed and incubated at 55 °C for 10 min. A total of 350 µL of ethanol was then added to each sample, and the samples were bound, washed, and eluted according to the manufacturer’s protocol. Four miRNAs (SNPs) were selected using the NCBI human genome SNP database (dbSNP, http://www.ncbi.nlm.nih.gov/snp (accessed on 13 March 2019)). The SNPs miR-10aA>T (rs3809783), miR-30cA>G (rs113749278), miR-181aT>C (rs16927589), and miR-499bA>G (rs37464444) are either mature-form (rs3746444, rs-formnp8978) or pri-form (rs3809783, rs16927589). miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G were genotyped according to TaqMan® SNP Genotyping Assays system (Applied Biosystems, Foster City, CA, USA). Based on the intensity of fluorescence signals of FAM and VIC, samples were automatically classified into one of three groups corresponding to the genotypes AA, AG, or TT of miR-10aA>T; AA, AG, or GG of miR-30cA>G; TT, TC, or CC of miR-181aT>C; and AA, AG, or GG of miR-499bA>G. The basic principle of the assay is as follows: when the allele-specific probe is fully hybridized to the template DNA, Taq polymerase cleaves the reporter dye, leading to fluorescence emission. However, if a single base mismatch exists between the probe and template DNA, hybridization is inefficient, and reporter dye fluorescence is thus reduced. The sequences of the SNPs were as follows: miR-10aA>T, CTCTT ATTTTTCCAG AAGAAAAAAA[A/T]ATATATATAT GTATATGTAG TATTT; miR-30cA>G, TACTTTCCACAGCTG AGAGTGTAGG[A/G]DTGTTTACAGT ATCTGTCGCT CAGTG; miR-181aT>C, AAAAT AGCACAAAAT TATCCAATTG[T/C] GACAGTTCTT ATCACATTTC ACTTT; and miR-499bA>G, ATGTTTAACT CCTCTCCACG TGAAC[A/G]TCACAGCAAG TCTGTGCTGC TTCCC. Information regarding the miRNA probes was as follows: miR-10aA>T, wild type homozygous AA (VIC reaction & FAM no reaction), heterozygous AT (VIC reaction & FAM reaction), mutant homozygous TT (VIC no reaction & FAM reaction); miR-30cA>G, wild homozygous AA (VIC reaction & FAM no reaction), heterozygous AG (VIC reaction & FAM reaction), mutant homozygous GG (VIC no reaction & FAM reaction); miR-181aT>C, wild homozygous TT (VIC reaction & FAM no reaction), heterozygous TC (VIC reaction & FAM reaction), mutant homozygous CC (VIC no reaction & FAM reaction); miR-499bA>G, wild homozygous AA (VIC reaction & FAM no reaction), heterozygous AG (VIC reaction & FAM reaction), mutant homozygous GG (VIC no reaction & FAM reaction).
Plasma PAI-1, total cholesterol, uric acid, and homocysteine levels were measured in participant blood samples. Plasma was separated by centrifugation of whole blood at 1000× g for 15 min. PAI-1 levels were determined using a human serpin E1/PAI-1 immunoassay (R&D Systems, Minneapolis, MN, USA). Uric acid and total cholesterol levels were measured using enzymatic colorimetric tests (Roche Diagnostics, GmbH, Mannheim, Germany). Homocysteine levels were measured using a fluorescence polarization immunoassay with an Abbott IMx analyzer (Abbott Laboratories, Abbott Park, IL, USA).
The significance of differences in the frequencies of the miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G SNPs between the control and patient groups were assessed using Fisher’s exact test and a logistic regression model. p-values were calculated using two-sided t-tests for continuous variables and chi-square tests for categorical variables. Allele frequencies were calculated to investigate the deviation from Hardy–Weinberg equilibrium. The genotype distribution of RPL patients and controls with ≥h or ≥o pregnancy loss was investigated. Odds ratios (ORs), adjusted odds ratios (AORs), and 95% confidence intervals (CIs) were used to examine the associations between various miRNA polymorphisms and RPL risk. Data are presented as the mean ± SD for continuous variables or a percentage for categorical variables. The results of the allele and genotype combination analysis were consistent with those derived from Fisher’s exact test during regression analysis. Statistical analyses were carried out using MedCalc software, version 12.1.4 (MedCalc Software bvba, Mariakerke, Belgium) or GraphPad Prism 4.0 software (GraphPad Software, Inc., San Diego, CA, USA). Logistic regression analysis was applied to data regarding baseline characteristics, genotype frequencies, genotype combinations, and allele combinations for quantitative traits shown in Table 2, Table 3, Table 4 and Table 5. The HAPSTAT program (v.3.0, www.bios.unc.edu/~lin/hapstat/ (accessed on 10 April 2018)), which exhibits a strong synergistic effect, was used to estimate the frequencies of polymorphic haplotypes. A p-value < 0.05 indicated statistical significance. HAPSTAT allows testing of haplotype (or allele combination) effects by maximizing the likelihood (from the observed data) that properly accounts for phase uncertainty and study design. False-positive discovery rate (FDR) correction was used to adjust multiple comparison tests and associations with FDR-adjusted p-values < 0.05 were considered statistically significant [30]. FDR calculation is also used for multiple hypotheses testing to correct for multiple comparisons. Multifactor dimensionality reduction (MDR) analysis was used to determine the best-model gene-gene interaction for RPL risk. The advantage of using MDR is that it overcomes the sample size limitations often encountered during logistic regression analysis in studies of high-level interactions. The MDR method consists of two main steps. First, the best combination of multi-factors is selected, and second, genotype combinations are classified into high- and low-risk groups [31]. We constructed all possible allelic combinations by MDR analysis to identify combinations with strong synergy. Allelic combinations for multiple loci were estimated using the expectation-maximization algorithm with SNPAlyze (v. 5.1; DYNACOM Co, Ltd., Yokohama, Japan), and those having frequencies < 1% were excluded from statistical analysis. We also applied multiple regression models to further explain the results of the allelic combination analysis. Genetic interaction analyses were performed using the open-source MDR software package (v.2.0), which is available at www.epistasis.org (accessed on 15 March 2018).
The pre-miRs (miR-10a, miR-30c, and miR-181a) and their flanking regions were amplified from human genomic DNA and cloned into the vector pcDNA3.1(−) (Invitrogen, Carlsbad, CA, USA). The primers used in the study included F: 5′-TGC GAA CTG GCT ACT TGA AA-3′, R: 5′-TTC CAA TAA AGC CTC CCT GA-3′ (miR-10a); F: 5′-GCA CCA TGT GTC ACA CAG GT-3′, R: 5′-CAA GTG TTG GGA AGA TGC TAT-3′ (miR-30c); and F: 5′-ACA TTT TCT CAG ACA TTC AT-3′, R: 5′-ATG TGA GAA AAC TGA GAC AC -3′ (miR-181a). For single-point mutations, we used an Intron Muta-direct kit (Intron, Seoul, Korea). The sequences of these three vectors were confirmed by direct sequencing, and the SNPs were the only differences detected. To generate the miRNAs target gene::luciferase reporter constructs, similar to the cloning vectors, fragments of the PAI-1 gene corresponding to the 3′-UTR region clone (OriGene, Rockville, MD, USA) were amplified and cloned into the pGL4.13-luciferase vector (Promega, Madison, WI, USA). The resulting cDNAs were PCR amplified using the following primers: forward 5′-CCC TGG GGA AAG ACG CCT T-3′ and reverse 5′-TTC GTA TTT ATT TAT TTT ATT TTT T-3′ with XbaI (TCTAGA)and FseI (GGCCGGCC) linker (New England Biolabs, Ipswich, MA, USA), and all constructs were verified by sequencing. Cells from a human endometrial cell line (Ishikawa) were plated at 1 × 106 cells per well in 6-well plates and transfected 24 h later using JetPRIME transfection reagent (Polyplus, France). Transfection reactions for miR-10a contained 500 ng of miR10a-A (in pcDNA3.1-) or 500 ng of miR-10a-T (in pcDNA3.1-) with 500 ng of 3′-UTR-PAI-1 in pGL4.13 and 200 ng of pGL4.75 (Renilla-normalization control); for miR-30c, reactions contained 500 ng of miR-30c-A (in pcDNA3.1-) or 500 ng of miR-30c-G (in pcDNA3.1-) with 500 ng of 3′-UTR-PAI-1 in pGL4.13 and 200 ng of pGL4.75 (Renilla-normalization control), for miR-181a2, reactions contained 500 ng of miR-181a-T (in pcDNA3.1-) or 500 ng of miR-181a-G (in pcDNA3.1-) with 500 ng of 3′-UTR-PAI-1 in pGL4.13 and 200 ng of pGL4.75 (Renilla-normalization control).
TRIzol reagent (Invitrogen, Waltham, MA, USA) was used to isolate total RNA from Ishikawa cells that were transfected with 2.5 μg of vector after 16 h. Total RNA was then reverse transcribed using an M-MLV reverse transcriptase PCR kit (Biofact, Co., Ltd., Daejeon, Korea) and random or oligo dT20 primers (Invitrogen, Waltham, Massachusetts, USA) in addition to specific primers for PAI-1 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Quantitative real-time PCR (qPCR) was performed as 20 μL reactions, containing each sequence-specific primer and quantitative PCR master mix (Solgent, Co., Ltd., Daejeon, Korea), using a Rotor-Gene 6000 real-time PCR system (Qiagen, Co., Ltd., Hilden, Germany). Expression levels were calculated according to the comparative threshold cycle (Ct) method using the formula 2−ΔΔCt. Primer sequences for amplification were as follows: has-miR-10a-pre forward: 5′-CCG AAT TTG TGT AAG GAA TTT TG-3′ and reverse 5′-AAG AGC GGA GTG TTT ATG TCA A-3′; has-miR-10a-mature forward: 5′-TAC CCT GTAG ATC CGA ATT T and reverse: universal primer (Qiagen Cat# 218193); has-miR-30c-pre forward: 5′-TGT GTA AAC ATC CTA CAC TCT CAG C-3′ and reverse: 5′-CCA TGG CAG AAG GAG TAA ACA-3′; has-miR-30c-mature forward: 5′-AAA CAT CCT ACA CTC TCA GC-3′ and reverse universal primer (Qiagen Cat# 218193); has-miR-181a-pre forward:5′-TAT CAG GCC AGC CTT CAG AG-3′ and reverse: 5′-AAT CCC AAA CTC ACC GAC AG-3′; miR-181a-mature forward:5′- TTC AAC GCT GTC GGT GAG TT-3′ and reverse: universal primer (Qiagen Cat# 218193); Human RNU6B (RNU6-2) forward:5′-ACG CAA ATT CGT GAA GCG TT-3′ and reverse universal primer (Qiagen Cat# 218193).
An online search was conducted to identify targets for miR-10a, miR-30c, miR-181a, and miR-499b using the TargetScan (http://www.targetscan.org (accessed on 21 May 2018)) and miRIAD databases (http://bmi.ana.med.uni-muenchen.de/miriad/ (accessed on 16 May 2018)). We used these databases to predict miRNAs that target overlapping regions of PAI-1 mRNA transcripts. Target mRNA sequences, particularly within the 3′-UTR, are often obtained from the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov/ (accessed on 26 September 2018)). We found that miR-30c, miR-10a, and miR-181a were predicted to be targets of the PAI-1 3′-UTR. Therefore, a luciferase reporter assay was used to evaluate the roles of miR-30c, miR-10a, and miR-181a in regulating the expression of target genes, as previously described. Briefly, wild-type pGL4.13-luciferase vector (Promega, Madison, WI, USA). constructs containing the 3′-UTRs of the PAI-1 gene were generated by amplifying the 3′-UTR region clone (OriGene, Rockville, MD, USA) and cloning the amplification products into the downstream region of the pGL4.13 vector (Promega, Madison, WI, USA) using the XbaI and FseI endonucleases (New England BioLabs, Ipswich, MA, USA). Positive clones were selected by sequence-specific PCR, restriction enzyme digestion, and DNA sequencing. Ishikawa cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Thermo Fisher Scientific, Inc. Waltham, Massachusetts, USA). All medium was supplemented with 10% fetal bovine serum (FBS) (Thermo Fisher Scientific, Inc. Waltham, Massachusetts, USA) and 1% penicillin/streptomycin (Thermo Fisher Scientific, Inc. Waltham, Massachusetts, USA). All cell lines were maintained in a CO2 incubator (5% CO2) at 37 °C. The Ishikawa cells used in this study were endometrial and are commonly used in RPL studies. Next, miR-10a, miR-30c, and miR-181a mimics (50 nM) were co-transfected into Ishikawa cells with 200 ng of the 3′-UTR of PAI-1 in pGL4.13 constructs using lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). After 16 h of incubation, the luciferase activity was measured using a dual-luciferase reporter assay system (Promega, Madison, WI, USA). Each transfection was performed as triplicates.
The characteristics of RPL patients and control subjects are summarized in Table 1. The mean age was approximately 33 years for both groups, and both groups were 100% female. PLT count, aPTT, and concentrations of E2 and LH were greater in RPL patients than in controls (p = 0.0007, p = 0.005, p = 0.001, and p = 0.011, respectively). There were no significant differences in age, BMI, uric acid level, or FSH level between the two groups.
Table 2 shows the distribution of genotypes in RPL patients with ≥3 or ≥4 pregnancy losses and control subjects. Significant differences in the miR-10a SNP were observed between the RPL and control groups and were significantly correlated with RPL prevalence. Consistently, the absence of these miRNA polymorphisms showed a negative correlation with RPL. The associations of these polymorphisms were very interesting in RPL patients because the miRNA polymorphisms were related to decreased RPL, but they were not associated with RPL risk (Table 2). In addition, the number of RPL patients with risk factors was very small. Therefore, the associations with RPL occurrence will require further investigation. miR-10aA>T (chr17:48579816, rs3809783), miR-30cA>G (chr6:71377017, rs113749278), miR-181aT>C (chr9:124692981, rs16927589), and miR-499bA>G (chr20: 34990400, rs37464444) were all in the miRNA mature-form (rs3746444, rs-formnp8978) or pri-form (rs3809783, rs16927589). The SNPs in miRNA genes, including pri-miRNAs, pre-miRNAs, and mature miRNAs, could potentially influence the processing and/or target selection of miRNAs. Since we selected four SNPs in pri-form or mature-form, we wanted to determine whether all these miRNAs could influence the expression and regulation of target genes. Based on the intensity of FAM and VIC fluorescence, samples were automatically classified into one of three groups corresponding to genotypes AA, AT, or TT of miR-10aA>T; AA, AT, or GG of miR-30cA>G; TT, TC, or CC of miR-181aT>C; and AA, AG, or GG of miR-499bA>G.
The miR-30cAG+GG genotype was associated with decreased risk of RPL for age < 33 years (odds ratio [OR] = 0.583; 95% confidence interval [CI] = 0.371–0.918; p = 0.022) (Table 3). However, the miR-181aTC+CC genotype was associated with increased risk of RPL for age < 33 years (OR = 1.677; 95% CI = 1.038–2.709; p = 0.035), and the miR-499bAG + GG genotype was associated with increased risk of RPL for age ≥ 33 years (OR = 1.631; 95% CI = 1.028–2.588; p = 0.038). The miR-10aAT+TT genotype was associated with increased risk of RPL for BMI ≥ 25 kg/m2 (OR = 2.840; 95% CI = 1.544–5.223; p = 0.001). The miR-499bAG + GG genotype was associated with increased risk of RPL for BMI <25 kg/m2 (OR = 1.456; 95% CI = 1.029–2.059; p = 0.034) and with increased risk of RPL for BMI ≥25 kg/m2 (OR = 2.284; 95% CI = 1.377–3.789; p = 0.001). The miR-181aTC + CC genotype was associated with increased risk of RPL for PLT count <255.62×103/μL) (OR = 1.779; 95% CI = 1.038–3.048; p = 0.036). Finally, the miR-30cAG+GG was associated with decreased risk of RPL for aPTT < 32.83 s (OR = 0.364; 95% CI = 0.185–0.717; p = 0.004).
The results of combined gene-genotype analyses are shown in Table 4. The miR-10a/miR-30c combined genotype AT/AG was associated with increased RPL risk (OR = 2.156; 95% CI = 1.120–4.151; p = 0.022). The miR-10aA>T/miR-181aT>C combined genotype AT/TT was associated with increased RPL risk (OR = 1.974; 95% CI = 1.065–3.658; p = 0.031). The miR-10aA>T/miR-499A>G combined genotype AT/AG was associated with increased RPL risk (OR = 2.195; 95% CI = 1.156–4.169; p = 0.016). The miR-30cA>G/miR-181aT>C combined genotype AG/TT was also associated with increased RPL risk (OR = 1.839; 95% CI = 1.054–3.210; p = 0.032). Similarly, increased RPL risk was associated with the miR-30cA>G/miR-499A>G combined genotypes AA/GG (OR = 4.324; 95% CI = 1.423–13.141; p = 0.010) and AG/AG (OR = 1.921; 95% CI = 1.145–3.224; p = 0.013). The miR-181aT>C/miR-499A>G combined genotype TT/GG was also associated with increased RPL risk (OR = 8.320; 95% CI = 1.043–66.384; p = 0.046). However, after false-discovery rate (FDR)-p correction, there were no significant differences between RPL patients and controls in the ORs for the combined genotypes, except for the miR-30cA>G/miR-499A>G combined genotypes AA/GG and AG/AG.
The results of allele combination analyses of miRNA polymorphisms in RPL patients and control subjects are shown in Table 5 and Supplementary Tables S2–S4. The allele combinations miR-10a/miR-30c/miR-181a/miR-499b A-T-G-G (OR = 1.952; 95% CI = 1.120–3.149; p = 0.006), A-C-A-G (OR = 2.343; 95% CI = 1.111–4.942; p = 0.026), A-C-G-A (OR = 2.136; 95% CI = 1.095–4.165; p = 0.028), T-T-G-A (OR = 0.455; 95% CI = 0.215–0.962; p = 0.044), and T-C-G-A (OR = 13.020; 95% CI = 0.739–229.300; p = 0.017) were associated with an increased risk of RPL. However, after FDR-p correction, there were no significant differences between RPL patients and controls in the ORs of the allele combinations, except for the A-T-G-G and T-C-G-A allele combinations.
The impact of SNPs on the interaction of miR-10aA>T, miR-30cA>G, miR-181aT>C, and miR-499bA>G on their targets was investigated by constructing various expression plasmids (pri-miR-10aA, pri-miR-10aG, pri-miR-30cA, pri-miR-30cG, pre-miR-181aT, pre-miR-181aG, pri-miR-499bA, and pri-miR-499bG) under control of the cytomegalovirus (CMV) promoter with either the major or minor allele. These plasmids were used in a dual luciferase assay performed with the 3′UTR of PAI-1, one of the predicted targets of miR-10a, miR-30c and miR181a, in Ishikawa human endometrial cells. A schematic diagram of a gene with a 3′-UTR of PAI-1 containing possible miR-10a and miR-30c binding sites in a conserved region is shown in Figure 1A,B. The luciferase activity of the 3′UTR of PAI-1 was significantly lower in pre-miR-10a having the A allele as compare to pre-miR-10a having the T allele (p < 0.05) (Figure 1C). Similarly, the luciferase activity of the 3′UTR of PAI-1 was significantly lower in the pre-miR-30c with the A allele as compared to pre-miR-30c with the G allele (p < 0.05) (Figure 1D).
Associations between miRNA polymorphisms and the levels of homocysteine, folate, total cholesterol, uric acid, blood urea nitrogen (BUN), estradiol (E2), thyroid-stimulating hormone (TSH), FSH, LH, prolactin, creatinine, platelets (PLT), as well as CD3+, CD4+, CD8+, CD19+, and CD56+ NK cells, in addition to the PT and aPTT were assessed by ordinal logistic regression analyses. We divided the risk factors into 10 grades and performed ordinal logistic regression using a proportional odds model. We found that the genotype frequency of miR-30cA>G was significantly associated with aPTT (AA: 32.46 ± 4.71, GG: 27.56 ± 3.59, p = 0.001), creatinine (AA: 1.19 ± 1.94, GG: 6.26 ± 3.71, p = 0.001), and E2 (AA: 1.19 ± 1.94, GG: 6.26 ± 3.71, p = 0.001). Levels of FSH differed significantly (p < 0.05) between the miR-30cA>G AA (mean ± SD, 32.36 ± 4.30 and 6.96 ± 4.29, respectively) and GG genotypes (30.49 ± 3.02 and 33.82 ± 55.85, respectively) (Table 6, Figure 2A,C,D). Additionally, levels of hematocrit (Hct) and total cholesterol (T. chol) differed significantly (p < 0.05) between the miR-30cAA and GG genotypes (36.65 ± 3.73 and 34.14 ± 4.49, 172.56 ± 64.85 and 38.15 ± 76.11, respectively). The miR-181aT>C genotype frequency was significantly associated with levels of creatinine (TT: 2.38±3.24, TC: 1.17±1.76, p = 0.011), Hcy (TT: 6.76 ± 2.01, CC: 9.98 ± 4.50, p = 0.001), LH (TT: 4.81 ± 2.74, CC: 4.20 ± 0.71, p = 0.038), PT (TT: 11.43 ± 1.14, CC: 10.20 ± 0.28, p = 0.048), and T. chol (TT: 136.96 ± 86.18, TC: 185.65 ± 76.23, p = 0.001). The miR-499bA>G genotype frequency was significantly associated with aPTT (TT: 31.20 ± 4.29, GG: 32.10 ± 4.18, p = 0.026) (Table 6, Figure 2B).
Increasing evidence suggests that miRNAs play critical roles in the pathophysiology of various reproductive disorders [14,15,32]. Here, we investigated whether four pre-miRNA SNPs (miR-10a, miR-30c, miR-181a, and miR-499b) were associated with the risk of RPL in a cohort of Korean women. Specifically, we focused on the genotypes and allele combination of the selected miRNA polymorphisms and aimed to determine how they affected the risk of RPL. Using a genotype-based analysis method, we found that the GG and dominant (AA vs. AG + GG) miR-499b genotypes were significantly more common in RPL patients (PL ≥ 3 and PL ≥ 4, p < 0.05) than control subjects. In allele combination analyses, the AA/GG and AG/AG genotypes of miR-30cA>G/miR-499A>G were significantly more common in RPL patients than in controls. As the activities of many genes are interconnected in complex conditions such as RPL, gene-gene interactions may affect gene-disease associations. The MDR method enables the detection of gene-gene interactions, regardless of the chromosomal locations of the genes [33]. We used a novel genotype-based MDR approach to examine the effects of potential interactions between different miRNAs on RPL risk. These results of these analyses, which examined the effects of four miRNA polymorphisms associated with RPL, suggested that gene-gene interactions involving these four miRNA polymorphisms also play roles in determining the risk of RPL. Allele combination MDR analyses indicated that the two combination conferred by the miR-10aA>T/miR-181aT>C/miR-30cA>G/miR-499A>G (A-T-G-G and T-C-G-A), the two combination conferred by the miR-10aA>T/miR-181aT>C/miR-30cA>G (T-T-A, T-C-G), the two combination conferred by the miR-10aA>T/miR-30cA>G/miR-499A>G allele combination (C-A-G, C-G-A), and the genotype conferred by the miR-10aA>T/miR-30cA>G allele combination (T-A) occur more frequently in patients with RPL than control subjects, suggesting a significant association with increased risk of RPL (all p < 0.05). In addition, the miR-10aA>T/miR-181aT>C/miR-30cA>G allele combination T-T-G and the miR-10aA>T/miR-30cA>G/miR-499A allele combination C-G-G were found to be less frequent in RPL patients than controls, suggesting these combinations exert a protective effect (all p < 0.05). SNPs that occur in miRNA genes, miRNA machinery genes, or miRNAs that target genes involved in miRNA synthesis or function could adversely affect downstream gene expression [34]. Several studies have provided evidence supporting the critical role of miRNAs in RPL [35]. A previous study demonstrated that miR-499 was associated with the transforming growth factor (TGF)-β signaling pathway [24]. Furthermore, the 3′-UTR of the TGF-β3 gene has been shown to contain a putative binding site for miR-30c (rs928508) (http://www.targetscan.org (accessed on 21 May 2018)), which targets the drug metabolism gene SULT1A1 [25]. Several TGF-β superfamily members perform critical functions in the female reproductive system. Specifically, these proteins regulate all processes of ovarian follicle development, including granulosa and theca cell proliferation, primordial follicle recruitment, gonadotropin receptor expression, ovulation, oocyte maturation, luteinization, and corpus luteum formation [36]. Additionally, the 3′-UTR of the prostaglandin F2 receptor inhibitor gene has been shown to contain a predicted binding target for miR-604 (http://www.targetscan.org (accessed on 21 May 2018)), and prostaglandin F2 is required for placenta retention [37]. Furthermore, the miR-10aA>T polymorphism has been associated with regulation of IL-6 expression [26], and a previous study reported abnormal IL-6 expression in both animal models and patients with recurrent spontaneous abortions [38]. An online search for miR-10a, miR-30c, miR-181a, and miR-499b targets using the Target Scan and miRIAD databases (http://bmi.ana.med.uni-muenchen.de/miriad/ (accessed on 21 May 2018)) returned many putative mRNA targets. Among these targets, we focused on PAI-1 for further functional analyses of miR-10a, miR-30c, and miR-181a because this gene has been shown to play several important roles in pregnancy and infertility [27]. PAI-1 is the primary inhibitor of plasminogen activators, including tPA and uPA. In the human placenta, PAI-1 is expressed in the extravillous interstitial and vascular trophoblasts. During implantation and placentation, PAI-1 inhibits extracellular matrix degradation, which thereby inhibits trophoblast invasion. We reviewed the literature regarding various reproductive diseases in which PAI-1 plays a role. Elevated PAI-1 levels have been detected in patients with RPL, preeclampsia, intrauterine growth restriction, gestational diabetes mellitus (GDM), endometriosis, and PCOS. Furthermore, both GDM and PCOS development have been reported to be related to the genetic role of the 4G/5G polymorphism in PAI-1. In general, elevated blood levels of PAI-1 are associated with an increased risk of infertility and poor pregnancy outcomes. In contrast, deficiency of PAI-1 results in transiently impaired placentation in mice [39], and deficiency of the PAI-1 gene is associated with abnormal bleeding after trauma or surgery in humans [40]. PAI-1 functions as a major inhibitor of fibrinolysis, and its overexpression leads to fibrin accumulation and placental insufficiency during pregnancy. PAI-1 acts as a major inhibitor of fibrinolysis, resulting in fibrin accumulation and insufficient placental formation due to overexpression. Previous reports also suggested that elevation of PAI-1 levels is the most frequent hemostasis-related abnormality associated with unexplained RPL [41]. Thus, increased expression of PAI-1 leading to inhibition of fibrinolysis is believed to be the main cause of RPL. To determine whether polymorphisms in miR-10a, miR-30c, and miR-181a affect target gene expression, we compared the expression levels of the 3′-UTR of PAI-1 harboring the different polymorphisms of miRNAs in Ishikawa human endometrial cells. Aberrant PAI-1 expression resulting from the expression of miR-10a with the A allele was significantly lower (p < 0.05) than aberrant PAI-1 expression resulting from the expression of miR-10a with the T allele. In addition, the expression of miR-30c with the A allele was significantly lower (p < 0.05) than expression of premature and mature miR-30c with the G allele. Expression of genotypes of miR-30cG as well as those of miR-10aT led to reduced expression of PAI-1 mRNA. These results suggest that SNPs in miR-30c and miR-10a regulate the expression of the PAI-1 gene. PAI-1-mediated inhibition of fibrinolysis and fibrin accumulation is currently believed to be the principal culprits for RPL; however, further studies are required to fully elucidate the underlying mechanisms. FSH is the primary gonadotropin responsible for regulating the progression of pregnancy [42]. Optimal levels of FSH, especially during the first few months of pregnancy, are critical for proper formation of the placenta [43]. Our clinical data indicated significant changes in FSH levels in RPL patients harboring the miR-30cA>G polymorphism. We, therefore, hypothesized that abnormal regulation of PAI-1 expression mediated by mutant miR-30c SNP results in aberrant FSH expression or disruption of the normal response to FSH. Imbalances in homocysteine and folate levels in particular are thought to contribute to low birth weight [44]. Specifically, higher homocysteine and lower folate concentrations during early pregnancy have been reported to be associated with lower placental weight and birth weight. However, we did not observe any associations between folate and homocysteine concentrations and placental weight. We found that the dominant miR-499b AG genotype (AA vs. AG + GG) was significantly more frequent in RPL patients (p < 0.05). Earlier studies used a global approach to identify and profile miRNA expression at important stages during the estrous cycle and found a role of miRNAs in ovulation. Additionally, one-way ANOVA analysis of variance of data from RPL patients (Table 6) revealed that in comparison with miR-30cAA, the miR-30cGG genotype was associated with significantly lower aPTT, E2 (pg/mL), Hct, and T. chol (mg/dL) and significantly higher creatinine (mg/dL) and FSH (mIU/mL). Compared with miR-181aTT, the miR-181aCC genotype was associated with significantly higher homocysteine levels, suggesting this genotype is associated with increased risk of RPL (p < 0.05). Compared with miR-181aTT, the miR-181aTC genotype was associated with significantly higher T. chol levels, suggesting this genotype is associated with increased risk of RPL (p < 0.05). However, in the case of creatinine levels, the miR-181aTC genotype was associated with significantly lower levels than the miR-181aTT genotype, indicating a protective effect, although the results were inconsistent with OR and therefore, the difference was not significant.
We investigated the relationship between various miRNA polymorphisms and the occurrence and risk of RPL. Several genotypes and allele combinations were positively correlated with RPL occurrence and unfavorable prognoses according to reproductive disease risk factors, including FSH, LH, and E2 levels. However, this study has several limitations. First, how the miRNA polymorphisms in the PAI-1 gene affect the development of RPL remains unclear. In addition to studies of PAI-1, future follow-up studies of other RPL-related genes and the miR-10a and miR-30c targets are planned, particularly studies of the role of genes related to the TGF-β signaling pathway. As TGF-β regulates cell proliferation, apoptosis, and homeostasis, it plays a critical role in regulating the progression of pregnancy. Second, the control subjects in our study were not completely healthy because some of them had sought medical attention for other issues. Our experience shows that recruiting healthy participants through imaging and laboratory testing results in significantly reduced enrollment rates. However, enrollment of participants without imaging and laboratory testing can introduce another challenge to risk factor assessment. Lastly, the study population was restricted to Korean patients. Although the results of our study provide the first evidence suggesting that miRNA polymorphisms in the PAI-1 gene may serve as diagnostic and prognostic biomarkers for RPL, a prospective study involving a larger cohort of patients is warranted to validate these findings. A genome-wide analysis (using transcriptome-seq and miRNA-seq) is needed to identify the primary target genes, particularly the common genes regulated by these miRNAs. Determining the expression of these genes in the relevant gene-miRNA networks would provide stronger evidence in support of the results of the present research. | true | true | true |
|
PMC9598526 | Jing Wang,Xifang Yang,Xueliang Wang,Wanhe Wang | Recent Advances in CRISPR/Cas-Based Biosensors for Protein Detection | 28-09-2022 | CRISPR,protein,detection,biosensors | CRISPR is an acquired immune system found in prokaryotes that can accurately recognize and cleave foreign nucleic acids, and has been widely explored for gene editing and biosensing. In the past, CRISPR/Cas-based biosensors were mainly applied to detect nucleic acids in the field of biosensing, and their applications for the detection of other types of analytes were usually overlooked such as small molecules and disease-related proteins. The recent work shows that CRISPR/Cas biosensors not only provide a new tool for protein analysis, but also improve the sensitivity and specificity of protein detections. However, it lacks the latest review to summarize CRISPR/Cas-based biosensors for protein detection and elucidate their mechanisms of action, hindering the development of superior biosensors for proteins. In this review, we summarized CRISPR/Cas-based biosensors for protein detection based on their mechanism of action in three aspects: antibody-assisted CRISPR/Cas-based protein detection, aptamer-assisted CRISPR/Cas-based protein detection, and miscellaneous CRISPR/Cas-based methods for protein detection, respectively. Moreover, the prospects and challenges for CRISPR/Cas-based biosensors for protein detection are also discussed. | Recent Advances in CRISPR/Cas-Based Biosensors for Protein Detection
CRISPR is an acquired immune system found in prokaryotes that can accurately recognize and cleave foreign nucleic acids, and has been widely explored for gene editing and biosensing. In the past, CRISPR/Cas-based biosensors were mainly applied to detect nucleic acids in the field of biosensing, and their applications for the detection of other types of analytes were usually overlooked such as small molecules and disease-related proteins. The recent work shows that CRISPR/Cas biosensors not only provide a new tool for protein analysis, but also improve the sensitivity and specificity of protein detections. However, it lacks the latest review to summarize CRISPR/Cas-based biosensors for protein detection and elucidate their mechanisms of action, hindering the development of superior biosensors for proteins. In this review, we summarized CRISPR/Cas-based biosensors for protein detection based on their mechanism of action in three aspects: antibody-assisted CRISPR/Cas-based protein detection, aptamer-assisted CRISPR/Cas-based protein detection, and miscellaneous CRISPR/Cas-based methods for protein detection, respectively. Moreover, the prospects and challenges for CRISPR/Cas-based biosensors for protein detection are also discussed.
Protein is an important type of disease biomarker for the early diagnosis of diseases, monitoring treatment process and outcome, and assessing prognosis [1]. Various protein detection methods have been established including immunoassay [2], biological mass spectrometry [3,4], fluorescence spectrometry [5], and electrical and electrochemical methods [6]. In particular, the immunoassays based on enzyme-linked immunosorbent assay (ELISA) [2,7] and chemiluminescence immunoassay (CLIA) [8] are currently the most commonly used protein assays, in which ELISA serves as the gold standard for protein detection in the fields of clinical diagnosis and biosafety [9]. However, the levels of protein biomarkers in clinical samples are generally very low, while a large amount of matrix interference exists in samples [10], so ELISA assays in clinical use are generally limited by their sensitivity, reliability, and specificity [11]. Therefore, it is urgent to develop new methods for the rapid, sensitive, portable, and highly specific detection of protein biomarkers. The CRISPR/Cas system is an adaptive immune system that originated from prokaryotes consisting of CRISPR sequences (Clustered Regularly Interspaced Short Palindromic Repeats) and proximity CRISPR-associated protein (Cas proteins), which can effectively and accurately identify and cleave foreign nucleic acids. This results in silencing their expression, and maintaining the stability of its genetic system through three stages of adaptation, recognition, and interference, thus effectively defending against foreign genes (e.g., phages and exogenous plasmids) [12,13]. In 1987, the CRISPR system was first discovered [14], which was subsequently coined as the CRISPR system in 2002 [15]. The CRISPR/Cas system has the advantages of programmability, specificity, sensitivity, and single-base resolution for nucleic acid recognition, and is now widely used in the biomedical fields [16,17,18]. It is generally divided into two types according to the structure of Cas proteins: Class I Cas proteins are effector complexes composed of multiple subunits; and class II Cas proteins are single effector proteins including Cas9, Cas12a, Cas13a, and Cas14 systems [19]. In recent years, some class II Cas proteins have been found to exhibit excellent signal amplification to neighboring non-target ssDNA or RNA with high non-specific cleavage efficiency [20,21]. For example, the Cas12a protein can specifically recognize the Protospacer Adjacent Motif (PAM) sequence of target DNA under the guidance of CRISPR RNA (crRNA) through forming a Cas12a/crRNA/DNA ternary complex, exhibiting non-specific cleavage activity to nearby ssDNA (trans cleavage activity) [22,23]. Unlike Cas12a, Cas13a is an RNA-mediated RNA endonuclease containing two HEPN structural domains that specifically recognize and cleave target RNA under the guidance of crRNA [24,25], which is able to indiscriminately cleave nearby RNA [26]. Due to the characteristics above-mentioned, CRISPR/Cas systems have been explored in the field of nucleic acid detection [27,28]. In 2017, Zhang’s team reported the seminal work of a CRISPR/Cas13a-based ultra-sensitive and specific method for the rapid detection of DNA and RNA, called Specific High-Sensitivity Enzymatic Reporter Unlocking (SHERLOCK), by combining the trans cleavage activity of the Cas13a protein with a fluorescently dual-labeled signal reporter [29]. Later on, Doudna’s team also conducted a landmark work of dubbed DNA Endonuclease Targeted CRISPR Trans Reporter (DETECTR) based on the CRISPR/Cas12a system, enabling nucleic acid detection at the attomolar level and the detection of two types of human papillomavirus (HPV), 16 and 18, in clinical samples [30]. Now, CRISPR technology has achieved remarkable success in the field of molecular diagnosis [31], which has recently become a hot topic in COVID-19 diagnosis [32,33]. Moreover, CRISPR diagnostic has been expanded for the detection of other substances such as ions [34,35], proteins [36], small molecules [37,38], etc., largely compensating for the limitations of traditional molecular diagnosis technology. For proteins, as the CRISPR/Cas system is only capable of recognizing and cleaving nucleic acid sequences, protein targets cannot directly activate the trans cleavage activity of Cas endonuclease, so it needs to convert the information of the protein molecule into the detectable nucleic acid signal that can be responded by CRISPR/Cas systems. To date, initial efforts have been made for the detection of protein biomarkers based on CRISPR/Cas12a systems due to the desirable characteristics of CRISPR diagnostics. However, most of the currently published reviews are about the research progress of CRISPR/Cas system applications in gene editing [39,40], gene therapy [41], bioimaging [42,43], pathogen diagnosis [44,45], and nucleic acid detection [46,47,48]. Although there are few reviews describing CRISPR-based biosensors for non-nucleic-acids, the protein detection section is not the main part, lacking a comprehensive and systematic summary on CRISPR-based protein detection [48,49,50]. A comprehensive literature review on the application of CRISPR/Cas systems in protein detection can provide a better understanding of CRISPR/Cas system applications in biosensing. Therefore, in this review, we present recent advances in CRISPR-based biosensors for protein analysis, which can be divided into three aspects based on signal conversion: antibody-assisted CRISPR/Cas-based protein detection, aptamer-assisted CRISPR/Cas-based protein detection, and miscellaneous CRISPR/Cas-based methods for protein detection, respectively (Scheme 1). We introduce their applications in protein detection with recent examples, and discuss their advantages, significance, and drawbacks. Finally, their challenges and potential for future applications are also discussed.
The antibody is a major recognition biomolecule with a symmetrical structure of two heavy chains (H chain) and two light chains (L chain) connected by disulfide and non-covalent bonds, capable of specifically recognizing and binding to antigenic determinants on the surface of the target proteins with high affinity and specificity [51,52]. The structure of the entire antibody molecule can be divided into two parts: the variable region (V region) and the constant region (C region) [53]. Antibodies are routinely used as biological recognition elements for proteins in immunosensors, which are currently the most important and widely used biosensors for protein detection [54]. The traditional ELISA method is well-recognized as the gold standard for protein detection [55], and its detection principle is based on the formation of a sandwich antibody–antigen–antibody structure, in which the enzyme (usually horseradish peroxidase (HRP)) labeled on the antibody induces the enzymatic signal amplification for measuring the concentration of targets. However, it is still not sensitive enough for the rapid detection of ultralow concentrations of protein biomarkers [56,57]. Moreover, the labeling of enzymes to antibodies usually requires complicated chemical modification and purification, easily resulting in the degradation of enzymes or antibodies. On the other hand, due to the high programmability, trans cleavage activity, and specificity of CRISPR systems with excellent signal amplification [58,59], the CRISPR/Cas systems have been combined with immunoassays for protein detection with significantly improved sensitivity. It was found that most of them were based on antibody–antigen–antibody type sandwich assays, and one was based on antigen–antibody recognition with the proximity CRISPR/Cas12a assay (Table 1). Li et al. reported a universal CRISPR-based immunosignaling enhancer called CRUISE, which constructed an antibody-ssDNA (Abs-ssDNA) through streptavidin-biotin binding to a biotinylated single-stranded DNA (ssDNA) [60]. Abs-ssDNA can act as a primary antibody to directly capture the target protein, and also indirectly as a secondary antibody to recognize the Fc fragment of the antibody used, forming a typical sandwich structure. Both methods can be used after washing away the unbound Abs-ssDNA, and the bound Abs-ssDNA activates the trans cleavage activity of CRISPR/Cas12a, cleaving the reporter molecule for the generation of a fluorescence signal. Secondary antibodies were integrated into a variety of different immunoassays to relieve the need of redundant recognition elements other than antibodies, providing three orders of magnitude higher sensitivity than traditional ELISA methods for IFN-γ detection. However, the CRUISE system has limitations similar to those of traditional ELISA methods such as the non-specific binding of Abs-ssDNA couplers that can reduce the analytical performance of the system and require better blocking strategies. Careful optimization of the antibody capture fixation method for 96-well plates is required to achieve higher sensitivity and specificity. Based on the classical sandwich-type sandwich structure, Chen et al. proposed a CRISPR/Cas13a signal amplification correlation immunosorbent assay called CLISA by designing biotinylated dsDNA containing a T7 promoter sequence instead of the traditional enzyme used for signal output (generally horseradish peroxidase) (Figure 1a) [61]. The presence of the target captures the secondary antibody on the capture antibody attached on the 96-well plate. The secondary antibody subsequently ligates the biotinylated dsDNA through biotin–streptavidin interaction, while the captured dsDNA is transcripted to generate a large amount of trigger RNA under T7 RNA polymerase. The trigger RNA activates CRISPR/Cas13a system under the assistance of crRNA along with the generation of fluorescence for protein detection. The CLISA detected human interleukin 6 (IL-6) with a limit of detection (LOD) of 45.81 fg/mL (2.29 fM) and human vascular endothelial growth factor (VEGF) with a LOD of 32.27 fg/mL (0.81 fM). It should be noted that a strict RNase-free environment is required for CLISA due to the use of RNA as the signal output. To improve the detection sensitivity, Lee and coworkers introduced antibody–DNA barcode conjugates with multiple Cas12a recognition sites into the conventional sandwich assay system using the affinity of biotin–streptavidin [62]. The detection signal was doubly amplified by increasing the number of Cas12a recognition sites on the DNA barcodes and the trans-cleavage activity of the CRISPR system. The assay achieved the detection of chemokine ligand 9 (CXCL9) in urine without PCR amplification, displaying a LOD of 14 pg/mL, which was seven times higher than that of the conventional ELISA method. The authors successfully evaluated CXCL9 protein in the urine of 11 kidney transplant patients with a 100% detection rate using this method, providing a potential tool for the non-invasive clinical diagnosis of kidney transplant rejection. To improve the sensitivity and binding specificity, Li et al. developed a universal proximity CRISPR/Cas12a assay by cleverly designing two target-specific primers with different lengths, P1 and P2, which were modified with affinity ligands that bind to different antigenic epitopes of the same antibody (Figure 1b) [63]. When the target is present, P1 and P2 are in proximity to each other. The primer extension reaction is triggered to generate a stable dsDNA containing the PAM sites, while the P1 is cleaved off using a nicking nuclease during the extension, which serves as the crRNA of the CRISPR/Cas12a system, to activate the trans cleavage activity of Cas12a, thus generating fluorescence and enabling the detection of the antibody at low concentrations of 1 pM. After further optimization, Li and coworkers reported that an improved iPCCA assay system achieved a LOD of IL-6 as low as 100 fM. This method can be applied in homogeneous solutions while maintaining detection sensitivity and does not require complex fixation and washing steps.
Due to the large molecular mass, high immunogenicity, and batch-to-batch variations of antibodies, the reliability and repeatability of antibody-based CRISPR biosensors may vary for each test, limiting their applications in protein detection. In the last decades, nucleic acid aptamers (simplified aptamers) have received the widespread attention of scientists, thanks to their excellent performance in sensing platforms, low cost, and comparable sensitivity [64]. Aptamers are synthetic functionalized single-stranded oligonucleotide sequences (DNA and RNA), also known as chemical antibodies, which are specific nucleic acid sequences and have three-dimensional structures, allowing them to bind target molecules with high affinity and specificity [65,66]. In contrast to antibodies, aptamers show the advantages of low immunogenicity, low preparation cost, long-term storage, ease of modifications, high stability, insensitivity to temperature, small size, no inter-batch variation, easy combination with nucleic acid signal amplifications, and applicability to a wide range of targets [67,68]. Aptamers are generally selected from nucleic acid molecular libraries by the Systematic Evolution of Ligands by EXponential enrichment (SELEX) [69,70], which was originally proposed by Tuerk and Ellington in 1990 [71]. The SELEX technology has successfully identified various aptamers for a range of proteins, which are widely used in cancer diagnosis, bioimaging, and therapy [72,73].
As the level of proteins is low in clinical samples, it usually needs to amplify the target proteins to effectively detect proteins. Most of the reported methods for protein amplifications depend on isothermal amplification such as hybrid chain reaction (HCR) [74], strand displacement amplification (SDA), [75,76], etc. In addition, traditional polymerase chain reaction (PCR) is also applied [77,78] due to its high sensitivity. Further powered by the amplification function of the CRISPR/Cas system, the integration of nucleic acid amplifications to CRISPR/Cas systems can achieve much improved sensitivity (Table 2). Prostate-specific antigen (PSA) is a serine protease produced by prostate epithelial cells, and its level is generally very low in normal human serum, but is abnormally high in the serum of prostate cancer patients [79,80]. Therefore, PSA is the most important prostate cancer biomarker, where its diagnostic specificity can reach more than 90% [81]. The Wang group developed a nicking enzyme-free SDA-assisted CRISPR/Cas12a-based colorimetric method for the detection of PSA with a LOD of 0.030 ng/mL (Figure 2a) [82]. When the PSA target is present, the released ssDNA opens the hairpin structure of the HP to release complementary ssDNA, triggering a nicking enzyme-free SDA reaction. The generated dsDNA serves as an activator of the CRISPR/Cas12a system to activate the trans cleavage activity of the Cas12a endonuclease, which non-specifically cleaves the nearby AuNP-linker probe. This allows the AuNPs to change from an aggregated purple state to a dispersed red state, which is colorimetrically determined, along with a visual readout. In this work, the exonuclease polymerase is harnessed for releasing cDNA from HP-cDNA during SDA, along with triggering the next SDA cycle, this unique design renders the biosensor simpler and more convenient for clinical testing. The same group also reported a colorimetric assay for serum PSA using the nonenzymatic and isothermal properties of HCR to convert serum PSA into nucleic acid products [83]. The presence of PSA triggers HCR amplification to produce dsDNA containing multiple PAM sites recognized by Cas12a, activating Cas12a’s trans cleavage activity, which nonspecifically cleaves the DNA–AuNP probe pairs along with a colorimetric signal. This strategy enables the sensitive and selective detection of PSA with a LOD of 0.10 ng/mL in both the spiked and clinical samples. Ultrasensitive detection of tumor-derived extracellular vesicles (TEVs) is key for the prognosis and diagnosis of cancers [84,85]. Li et al. developed a PCR-powered-CRISPR/Cas12a assay, which consists of three parts: aptamer recognition, PCR amplification, and CRISPR/Cas12a detection (Figure 2b) [86]. The aptamers for membrane proteins were coated on microtiter plates, which can specifically recognize and bind to extracellular vesicle surface membrane proteins for the formation of a sandwich-type complex. After washing away the unbound aptamers, the bound aptamers were amplified by PCR to generate a large amount of dsDNA, which activates the trans cleavage activity of Cas12a, enabling the detection of CD109+ and EGFR+ TEV at a concentration as low as 100 particles/mL. Moreover, the linear range spans six orders of magnitude (102–108 particles/mL), which is sufficient to detect TEVs in low volume (50 μL) samples. However, this method uses PCR amplification strategy, and its thermal cycling process requires complex instrument control, which limits its clinical application. Moreover, high temperature during the thermal cycling process may denature the proteins, affecting the sensing performance. Zhao et al. also reported an HCR amplified CRISPR/Cas12a-based biosensor, named AID-Cas, for the wash-free detection of EVs in the concentration range of 102–106 particles/μL [87]. The CD63 aptamer structural domain contained on the variant probe specifically recognizes and binds CD63+ TEVs, triggering double-loop HCR amplification. The amplified dsDNA contains a T7 promoter recognition sequence, which can be recognized by T7 RNA polymerase and transcribed to a large amount of RNA, serving as the crRNA of the CRISPR/Cas12a system. This method enables the quantitative detection of TEVs in the cell culture supernatants and clinical samples. However, the free CD63 protein from ruptured EVs or cells may interfere with the assay results. Similarly, Xing et al. developed an apta-HCR-CRISPR assay for the ultra-sensitive quantification of TEV surface proteins, which uses HCR to amplify the TEV surface proteins based on the corresponding aptamers, generating dsDNA containing multiple PAM sites for activating the trans cleavage activity of Cas12a in the presence of crRNA [88] (Figure 2c). This method can directly be used for the clinical analysis of circulating TEVs in 50 μL serum, achieving TEV detection at a concentration as low as 102 particles/µL in complicated biological samples.
The amplification of target proteins often requires complicated pre-detection processing, which often leads to problems such as non-specific amplification or reagent contamination [89,90]. To overcome these problems, a series of amplification-free sensing strategies have been proposed for protein analysis such as multiple activator dsDNA, the use of ssDNA blockers, the combination with electrochemical signal amplification, multiple Cas recognition sites, and other method-coupled CRISPR/Cas techniques (Table 3). Zhao et al. designed a dual aptamer sensor to implement a multi-trigger dsDNA tandem binding CRISPR/Cas12a system for the PCR-free detection of the SARS-CoV-2 antigenic nucleocapsid protein (Np) (Figure 3a), in which a hybrid DNA containing Cas12a-triggered dsDNA (HyDNA) modified with two aptamers, A48 and A61, was used to recognize different epitopes of Np [91]. When Np is present, the aptamers release two HyDNA activators from the hybrid DNA, subsequently activating Cas12a’s trans cleavage activity for non-specifically cleaving a nearby fluorescently labeled ssDNA probe along with the generation of the fluorescence signal. This multi-trigger dsDNA tandem element may be able to serve as a versatile tool for implementing highly sensitive CRISPR biosensors. Similarly, Li et al. used a DNA walker to amplify “one-to-many” by embedding the target’s aptamer in the locked strand and then hybridizing it with the walking strand [92] (Figure 3b). The walker is efficiently driven by using a nicking endonuclease as the energy supply. When the target is present, the walker can compete for the release of the walking strand to generate multiple activators, thus activating the trans cleavage activity of Cas12a to generate a fluorescent signal. To validate the performance of the method in real samples, the authors applied it to the detection of inactivated SARS-CoV-2 in the saliva and serum spiked samples, with a positive detection rate of 100%. The LOD for carcinoembryonic antigen is 0.32 pg/mL and features a fluorescent reporter gene loaded onto a biochip coated with photonic crystals (PC) and excited by a mini-type portable blue light, allowing the results to be observed with only a smartphone without the need for other sophisticated imaging tools. Zhao et al. immobilized biotinylated ssDNA partially bound to the designed aptamer-dsDNA complex on streptavidin-coated magnetic beads (MBs), in which the target can bind to the aptamer to release the dsDNA [93] (Figure 3c). The released dsDNA is used as an activator of the CRISPR/Cas12a system, triggering the trans cleavage activity of Cas12a, and thus non-specifically cleaving the fluorescently labeled ssDNA signal probe with a fluorescent signal output. This enabled the analysis of the tumor biomarker alpha-fetoprotein (AFP) in less than 20 min with a LOD as low as 0.07 fM/L, while the quantitative analysis of cocaine was at a LOD of 0.34 μmol/L. This highly modular biosensing platform has great potential for the detection of other analytes. Under the condition of very low magnesium ion (Mg2+) concentration, Liu et al. broadened the biosensing application of CRISPR/Cas13a by introducing an ssDNA blocker modified with an aptamer at the end to program the trans cleavage activity of CRISPR/Cas13a [94] (Figure 3d). The ssDNA blocker binds to crRNA and blocks the conformational change of crRNA to inhibit the trans cleavage activity of Cas13a. The presence induces the ssDNA blocker to release crRNA, restoring the Cas13a’s trans cleavage activity along with enhanced fluorescence signal. Moreover, this strategy is applicable to the detection of analytes that can bind to the ssDNA blocker to release crRNA. In addition, it was found that the Mg2+ concentration plays an important role in the activity of the Cas13a protein. If the Cas13a protein is highly active, its enzyme activity will not be easy to block, resulting in high fluorescence background, so a suitable Mg2+ concentration is demanded to block its activity effectively. Controlling the structure of crRNA by a simple ssDNA blocker to regulate the trans cleavage activity of Cas13a offers new opportunities for the development of CRISPR/Cas13a biosensors. In 2019, Liu’s team combined electrochemistry with the CRISPR system to develop an aptamer-based E-CRISPR tandem technology for protein detection for the first time [95] (Figure 3e). The authors validated the detection performance of the E-CRISPR electrochemical biosensor using the growth factor beta 1 (TGF-β1) protein by square wave voltammetry (SWV). The ssDNA aptamer also serves as the template for activating the trans cleavage activity of Cas12a, so the aptamers activate the CRISPR/Cas12a system without the addition of TGF-β1, cleaving the ssDNA reporter (methylene blue-labeled) with no electrochemical signal output. In contrast, the presence of TGF-β1 weakens the trans cleavage activity with more intact ssDNA reporters and a stronger electrochemical signal. The linear range was up to three orders of magnitude, with a LOD of 0.2 nM, while the detection was completed in 60 min. Very recently, Yuan et al. reported the CRISPR/Cas12a coupled voltage enrichment by coupling electrochemical and CRISPR systems [96]. The authors designed two aptamers, AptVEGF-HBD-T18-MB and AptVEGF-RBD, which can specifically recognize the HBD and RBD domains of vascular endothelial growth factor (VEGF), respectively. Among them, AptVEGF-HBD-Tx-MB consists of a thiol group at the 5′ end and a different internal T (Tx) site or MB tag at the 3′ end of the aptamer, which is covalently modified to AuNPs@Ti3C2TXMxene/GCE through Au–S bonding. When the target is present, AptVEGF-HBD-Tx-MB is recognized and bound to it, bringing the MB close to the electrode surface, improving the electron transfer efficiency, and generating a “signal-on” response; when the target is not present, the MB is away from the electrode surface, resulting in the “signal-off” response; when a positive voltage of 0.4 V was applied, the negatively charged MB groups were rapidly attracted to the electrode surface, resulting in a stronger current signal, resulting in a “signal superconducting” response. This strategy goes through a “signal on–off–on” sandwich-type mode for the detection of VEGF rather than a complicated target amplification step to enrich the cleaved signal probe. Converting the “signal-off” of CRISPR/Cas12a cleavage to “signal super on” further improves the current response, thereby simplifying the routine detection process and amplifying the electrochemical signal. The linear range of VEGF detection was from 1 pM to 10 μM in the serum samples, with a LOD of 0.33 pM. Electrochemiluminescence (ECL) is chemiluminescence that originates from the electron transfer between species generated on the surface of electrodes [97,98], which simultaneously has the dual advantages of electrochemical analysis and chemiluminescence such as high sensitivity, good reliability, simple operation, and fast analysis process [99,100], so it has been widely used in the biomedical field [101,102], food safety assessment [103,104], environment monitoring [105,106], and other fields of biomolecule detection [107,108]. Based on the merits of ECL, Liu et al. combined the advantages of spherical nucleic acids with CRISPR technology, in which a Y-shaped DNA structure constructed from helper DNA (A1), prostate cancer biomarker α-methylacyl coenzyme A racemase (AMACR) adaptors, and DNA activators are loaded onto gold nanoparticle-modified Fe3O4 magnetic beads (Au@Fe3O4MBs) [36]. Y-SNA serves as a target transducer to convert the protein signal into the programmable nucleic acid signal, while 1-pyrenecarboxaldehyde (Pyc) as a nanoemitter is embedded in magnetic mesoporous silica nanoparticles (MMSNs). Meanwhile, silver nanoparticles (AgNPs) serve as a co-reaction gas pedal to synergize with Pyc, and the synthesized AgNP-Pyc@MMSNs nanomaterial has a strong and stable ECL signal. The presence of the target protein induces the release of the DNA activator of Cas12a, activating the trans cleavage activity of Cas12a and thus non-specifically cleaving the ferrocene-labeled quenching probe (QP) in its vicinity with the ECL signal output. The designed ECL biosensor was used to determine AMACR from 10 ng/mL to 100 μg/mL, with a LOD of 15.8 pg/mL. However, the complicated pre-processing step causes difficulties in the large-scale production of biosensor-related reagents for further clinical applications, and the vulnerability to protease denaturation during the preparation process dampens the performance of the biosensors. Li et al. proposed a novel aptamer-based CRISPR/Cas12a immunoassay method called ALCIA, which established a link between non-nucleic acid targets and the CRISPR/Cas12a system by modifying the analyte-targeted aptamer (named Apt-acDNA) at the 5′ end of the activator DNA (acDNA), which can activate the trans cleavage activity of Cas12a upon target recognition, along with the fluorescence signal [109]. The authors designed dual aptamers based on two identical subunits of platelet-derived growth factor BB (PDGF-BB), where one aptamer was modified on the plate substrate to capture PDGF-BB and the other was modified at the 5′ end of acDNA to release the activator DNA. When PDGF-BB is present, it can form a sandwich-like structure of aptamer/PDGF-BB/Apt-acDNA for activating Cas12a. This assay detects PDGF-BB in the serum, urine, and saliva during a narrow range of 0–150 pM, with a LOD of 1.57 pM. The output signal of ALCIA can be adapted based on the actual needs. Moreover, its sensing principle is similar to ELISA and is highly compatible with traditional ELISA methods, which has the great potential for bioanalytical analysis and clinical testing. Similarly, Deng et al. developed a CRISPR/Cas12a-assisted fiber-optic immunosensor (CAFI) that could detect IFN-γ in the serum, urine, and saliva with a LOD of 1 fg/mL (58.8 aM) over a detection range of 1 fg/mL to 100 pg/mL [110]. By modifying biotinylated capture antibodies on the surface of antifouling glass fibers modified with silane-polyethylene glycol-biotin and streptavidin, an antibody–analyte–adaptor sandwich structure can be formed in the presence of the target. The CAFI assay system can be applied for other analytes such as insulin detection and analysis by simply changing the aptamer and capturing antibodies. However, its detection time takes 4 h, which hampers its on-site applications. Increasing the temperature of the reaction to 37 °C may be a feasible solution in reducing the detection time while maintaining sensitivity. Zhao et al. covalently modified aptamer and Cas12a target DNA activators on AuNPs to form sandwich-like structures of antibody–target–aptamers [111]. When the target is present, the sandwich-like structure forms, while the activators modified on AuNPs are captured for activating the trans-cleavage activity of Cas12a, thus cleaving ssDNA modified with a fluorophore (FAM) and a quencher (BHQ1) at both ends, along with an increasing fluorescence signal (Figure 3f). The reported Cas12a/crRNA-based nano-immunosorbent assay (Nano-CLISA) can determine carcinoembryonic antigen (CEA) and PSA in clinical samples with the LODs of 13.9 fg/mL and 5.6 fg/mL, respectively. AuNP-modified oligonucleotides to activate CRISPR/Cas12a can greatly enhance the sensitivity of the assay, which makes the assay 1000 times sensitive than conventional ELISA methods.
In addition to protein signal conversion methods based on well-recognized antibodies and aptamers, some other signal conversion methods have also been reported such as the use of PAM-free conditional DNA substrate (pcDNA), protection experiments with the help of nucleic acid exonuclease III, the use of protease-activated RNA polymerase, and the use of small molecule modification activator ssDNA (Table 4). The CRISPR/Cas system has become a powerful tool for live cell imaging, but its utility is limited to genomic loci and mRNA imaging in living cells [112,113]. On the other hand, the recognition of the CRISPR/Cas12a system dsDNA heavily depends on PAM sequences [114], which greatly limits the detection scope of targets. The design of DNA substrates using a universal response mechanism can expand the types of analytes. Inspired by the fact that the Cas12a/gRNA complex can recognize unwound DNA substrates without the restriction of PAM [115], the Nie team rationally constructed the unwound seed region and introduced a bubble structure in the seed region to make it unwind to overcome the CRISPR/Cas12a system’s PAM limitation (Figure 4a) [116]. By designing a pcDNA, the target recognizes and converts the corresponding pcDNA into a PAM-free dsDNA substrate (pDNA). pDNA activates the nuclease activity of Cas12a to nonspecifically cleave the surrounding fluorescent ssDNA signal probe for living cell imaging, and this PAM-free strategy was found to be suitable for adenosine triphosphate (ATP), miRNA, and telomerase. This biosensor can also be applied for the sensitive sensing of a wide range of biomolecules such as intracellular enzymes, small molecules, and microRNAs. The main limitation of this biosensor is that its reaction kinetic depends on the effective collision of reactants in the cytoplasm, which leads to reduced sensitivity and reproducibility due to the complicated biological environment inside the cells. Encapsulating the components of the sensing system into a restricted space by DNA technology or liquid–liquid phase separation may be an effective solution to alleviate this limitation. Cheng et al. also reported a AuNP-assisted CRISPR/Cas system for the visual detection of telomerase activity in three cases: positive (P), negative (N), and false-negative (FN) [117]. The authors designed telomeric repetitive sequence DNA and internal control crRNAs, crRNA1 and crRNA2, respectively. Both Cas12a/crRNA1 and Cas12a/crRNA2-mediated assays in the positive state keep AuNPs in the dispersed state. Cas12a/crRNA1-mediated assays in the negative state induce the cross-linking of AuNPs, and Cas12a/crRNA2-mediated assays ensure that the AuNPs remain dispersed. The false-negative state due to the PCR inhibitor or telomere repeat amplification protocol (TRAP) reagent errors allowed for both the Cas12a/crRNA1 and Cas12a/crRNA2-mediated analysis to induce the cross-linking of AuNPs. The platform was able to visually identify false-negative results caused by PCR inhibitor and TRAP reagent errors free of a complicated polyacrylamide gel electrophoresis (PAGE) process, significantly improving the accuracy of conventional TRAP. The authors also validated that the Cas9-mediated TL-LFA platform can also be used for accurate telomerase activity detection, which can be achieved within 15 min on a single test strip. Exonuclease III (ExoIII) recognizes flat-ended dsDNA and cleaves it from 3′ to 5′ to produce ssDNA with a 3′ protruding end [118]. An ExoIII-assisted Cas12a biosensing system is reported for the detection of transcription factors (TFs) based on which the activator dsDNA of Cas12a also contains the structural domain of TFs [119] (Figure 4b), in which TFs can bind to the activator to prohibit the degradation of the dsDNA by ExoIII. The intact dsDNA activator is thermally inactivated at 65 °C, which is further used to activate the trans cleavage activity of Cas12a. This method is applied for the detection of the nuclear factor-κB (NF-kB) p50 subunit with a LOD of 0.2 pM. The method has the potential to screen TF inhibitors and evaluate their biological activities. However, it should be noted that the method is limited to its long detection time and requires temperature control. Moreover, the assay performance is affected by the cellular nuclear protein extracts. By using protease-activatable RNA polymerases (denoted as PRs), Yang et al. transformed protein hydrolysis events into multiple programmable RNA sequences by in vitro transcription using PRs as transducers, and protease hydrolysis can activate RNA polymerase transcription to produce RNA, which serves as a guide RNA (gRNA) for activating the CRISPR/Cas12a system in the presence of template dsDNA, resulting in a corresponding fluorescent signal output [120] (Figure 4c). The authors combined protein hydrolysis-triggered signaling transcriptional events with the trans cleavage activity of the CRISPR/Cas system to achieve dual signal amplification. This strategy was used to detect protease biomarkers at the femtomolar level, with a LOD of 47.8 fM and 5.4 fM for thrombin and matrix metalloproteinase-2 (MMP-2), respectively. The sensitivity of the method was 5–6 orders of magnitude lower than the traditional peptide-based methods. This strategy extends the CRISPR/Cas system to the activity analysis of protease biomarkers, providing a new way for protease activity detection. Given the modularity of PR-Cas, the activity of other proteases can be assessed by simply replacing the PR module. Furthermore, the CRISPR/Cas system can also be extended for the detection of intermolecular interactions. Kim and coworkers developed a method for the rapid detection of protein/small molecule interactions based on the CRISPR/Cas system using a small molecule modified activator ssDNA (AD) that interacts with the target protein, and the interaction between the small molecule and the protein prevents AD from binding to crRNA, reducing the trans cleavage activity of Cas12a with decreasing fluorescence, which was used for the detection of streptavidin/biotin and antidigoxin/digoxin with the LODs of 0.03 nM and 0.09 nM, respectively, and this process was completed within 11 min (Figure 4d) [121]. In theory, this strategy can be used for the rapid detection of other protein–small molecule interactions, offering a new perspective on protein–small molecule interaction analysis and the screening of related modulators.
CRISPR-based biosensors have achieved huge success in nucleic acid analysis, but studies on the applications of CRISPR for protein detection are still relatively limited. Encouragingly, antibody-combined CRISPR/Cas biosensors have largely improved the LODs of protein detection, and expanded the detection range. It should be noted that most of the CRISPR-based biosensors for protein detection employ aptamers as signal recognition elements because of their superior integration and molecular properties, which easily combine with CRISPR/Cas systems for protein recognition, converting protein signals to nucleic acid signals with activated Cas and signal output. Moreover, with the rapid development of SELEX technology, more and more aptamers for proteins will be discovered [122,123], which will widely expand the applications of the aptamer-assisted CRISPR/Cas biosensors for protein detection. However, for further routine application and commercialization, the CRISPR/Cas-based protein detection system still faces a range of challenges. The most critical issue is how to efficiently convert the protein signal into a nucleic acid signal, thus activating the trans cleavage activity of Cas enzymes. To date, most of the methods reported so far depend on the antibodies and aptamers, with a few strategies coupled with other methods. These methods generally suffer from problems such as multi-step detection, ease of contamination, the need for specialized technicians, and reduced reliability in real samples. In addition, many CRISPR/Cas-based biosensors for protein detection still need a long detection time and sensitivity that do not fully meet the needs of clinical testing. To accelerate the practical application of CRISPR-based biosensors for protein analysis, it will encourage combining CRISPR/Cas systems with other advanced techniques. For example, the emergence of new technologies may better facilitate the development of CRISPR-based biosensing systems for protein detection such as bioinformatics, which could create easy access to predict and design gRNA/crRNA and target activators to improve the sensitivity and specificity of the sensing systems. Automation and high-throughput techniques can be integrated with the CRISPR/Cas system to develop biosensors that can rapidly screen large numbers of samples simultaneously and are easy to perform for protein analysis. Other portable devices (e.g., paper-based devices or microfluidic devices) may also be compatible with CRISPR-based protein sensing systems to meet the needs of clinical analysis. Moreover, as crRNA, PAM sequences, and Mg2+ are key to the sensing performance for this type of biosensor, it needs to carefully optimize the reaction system to achieve rapid quantitative protein detection, and improve detection sensitivity and reliability as well as to simplify the detection steps and reduce cost. By combining multiple Cas enzymes with different functions, it may be possible to achieve multiple assays simultaneously. Furthermore, the difficulty in detecting proteins reliably in complicated matrices may be addressed by introducing well-developed preprocessing methods (e.g., extraction, centrifugation, etc.) and encapsulating the components of the sensing system into a restricted space (e.g., DNA technology or liquid–liquid phase separation). Although the developed methods above only have one or two original targets, most of them have the potential to be extended to other proteins by simply changing the aptamer or antibody. Therefore, it is economically efficient to investigate their feasibility in other analytes, which will largely reduce the cost and accelerate their applications. Based on the rapid development of the technology, the deepening research on CRISPR/Cas systems, and the discovery of new CRISPR/Cas systems, we believe that the CRISPR/Cas technology will become one of the mainstream protein detection tools in the future, facilitating its rapid development in disease diagnosis, pathogen analysis, environmental assessment, and other fields. | true | true | true |
|
PMC9598558 | Thomas J. Kalantzakos,Luke E. Sebel,James Trussler,Travis B. Sullivan,Eric J. Burks,Carmen D. Sarita-Reyes,David Canes,Alireza Moinzadeh,Kimberly M. Rieger-Christ | MicroRNA Associated with the Invasive Phenotype in Clear Cell Renal Cell Carcinoma: Let-7c-5p Inhibits Proliferation, Migration, and Invasion by Targeting Insulin-like Growth Factor 1 Receptor | 28-09-2022 | clear cell renal cell cancer (ccRCC),microRNA,insulin like growth factor 1 receptor (IGF1R),proliferation,migration,invasion | Differential microRNA (miRNA) expression can portend clear cell renal cell carcinoma (ccRCC) progression. In a previous study, we identified a subset of dysregulated miRNA in small renal masses, pT1 ccRCC (≤5 cm) that are associated with an aggressive phenotype. The present study investigated miRNA expression in clinical stage I (cT1) tumors (≤5 cm), comparing pathologic stage I (pT1) tumors to those upstaged to pathologic stage 3 (pT3) after surgery following identification of renal vein invasion or invasion into adjacent fat tissue within Gerota’s fascia. Twenty cT1 tumors were examined in an miRNA screening, 10 pT1 and 10 pT3 tumors. The ccRCC cell lines 786-O and Caki-1 were used to assess the impact of let-7c-5p and its protein target insulin-like growth factor 1 receptor (IGF1R). Cells were transfected with pre-let-7c-5p and assessed through cell proliferation, migration, and invasion assays. IGF1R expression was evaluated through Simple Western, and interaction between let-7c-5p and IGF1R was confirmed via luciferase reporter assay. Screening identified 20 miRNA, including let-7c-5p, that were dysregulated between pT1 and pT3 upstaged tumors. This miRNA was also downregulated in our previous study of pT1 tumors that progressed to metastatic disease. Transfection of ccRCC cells with pre-let-7c-5p significantly inhibited proliferation, migration, invasion, and IGF1R expression. These findings suggest that miRNA dysregulation is involved in ccRCC progression, specifically through invasion, and that let-7c-5p downregulation contributes to the aggressiveness of small ccRCC tumors, in part, through its regulation of IGF1R. | MicroRNA Associated with the Invasive Phenotype in Clear Cell Renal Cell Carcinoma: Let-7c-5p Inhibits Proliferation, Migration, and Invasion by Targeting Insulin-like Growth Factor 1 Receptor
Differential microRNA (miRNA) expression can portend clear cell renal cell carcinoma (ccRCC) progression. In a previous study, we identified a subset of dysregulated miRNA in small renal masses, pT1 ccRCC (≤5 cm) that are associated with an aggressive phenotype. The present study investigated miRNA expression in clinical stage I (cT1) tumors (≤5 cm), comparing pathologic stage I (pT1) tumors to those upstaged to pathologic stage 3 (pT3) after surgery following identification of renal vein invasion or invasion into adjacent fat tissue within Gerota’s fascia. Twenty cT1 tumors were examined in an miRNA screening, 10 pT1 and 10 pT3 tumors. The ccRCC cell lines 786-O and Caki-1 were used to assess the impact of let-7c-5p and its protein target insulin-like growth factor 1 receptor (IGF1R). Cells were transfected with pre-let-7c-5p and assessed through cell proliferation, migration, and invasion assays. IGF1R expression was evaluated through Simple Western, and interaction between let-7c-5p and IGF1R was confirmed via luciferase reporter assay. Screening identified 20 miRNA, including let-7c-5p, that were dysregulated between pT1 and pT3 upstaged tumors. This miRNA was also downregulated in our previous study of pT1 tumors that progressed to metastatic disease. Transfection of ccRCC cells with pre-let-7c-5p significantly inhibited proliferation, migration, invasion, and IGF1R expression. These findings suggest that miRNA dysregulation is involved in ccRCC progression, specifically through invasion, and that let-7c-5p downregulation contributes to the aggressiveness of small ccRCC tumors, in part, through its regulation of IGF1R.
Renal cell carcinoma (RCC) is a heterogeneous group of cancer variants arising from a variety of specialized cells along the length of the nephron [1]. Clear cell renal cell carcinoma (ccRCC) is the most common histologic subtype of RCC, accounting for 75–80% of cases [1]. ccRCC is disproportionately represented within metastatic RCC, comprising over 90% of cases [2]. One area of particular interest centers on small renal masses, which have seen a disproportionate rise in incidence through the rise of cross-sectional imaging [3]. Clinical staging of incidentally discovered ccRCC tumors relies on radiographic size criteria and gross invasion of venous structures or within Gerota’s fascia [4]. However, microscopic renal vein and perirenal fat invasion is not adequately detected through imaging, prompting an upstaging of clinical stage I (cT1) tumors to pathologic T3a (pT3a) at the time of surgery [5]. This phenomenon has been identified in about 5% of patients and is associated with a substantial increase in mortality risk [5]. Recent studies have concentrated on the identification of novel prognostic indicators to improve treatment algorithms [5]. The potential of microRNA (miRNA) as an identifiable biomarker in ccRCC tumors has previously been demonstrated in many instances, including our previous study characterizing miRNA in pathologic stage I (pT1) tumors that progressed to metastatic disease [6]. miRNAs are small, highly conserved non-coding RNA molecules that bind to the three prime untranslated regions (3′UTR) of mRNA transcripts, silencing gene expression [7]. miRNAs exert their influence through targeting important signaling pathways and regulatory networks [8]. In addition to its prognostic value, miRNA dysregulation has been identified as a driver of metastasis and disease progression in ccRCC [9], and has been associated with early relapse after nephrectomy in RCC patients [10]. We previously investigated small pT1 ccRCC tumors (pT1 ≤ 5 cm) that ultimately progressed to metastatic disease [6]. One particular miRNA which corresponded to poorer cancer-specific survival was let-7c-5p, which was downregulated in aggressive small renal masses. The prognostic value of let-7c-5p expression in ccRCC and the conferral of aggressiveness through its downregulation in other cancers has been established. In paired ccRCC tissue specimens, let-7c was downregulated in ccRCC tissue compared to adjacent normal [11]. Expression of let-7c was also demonstrated to be significantly reduced in aggressive early metastatic ccRCC tumors compared with non-metastatic [9]. Additionally, elevated levels of let-7c-5p in the urine can be used to differentiate RCC patients from healthy controls [12]. Let-7c-5p has also been established as a tumor suppressor in a wide range of other cancers [13,14,15,16,17,18]. In non-small cell lung cancer, for example, a significant association was demonstrated between low let-7c-5p expression and metastasis, venous invasion, advanced TNM stages, and poor survival [17]. A miRNA exerts its influence on cellular phenotype through the specific protein targets that it interacts with. These interactions exist as part of a complex network, where each miRNA can silence hundreds of unique targets, while each protein is regulated by many miRNAs [19]. The role of a particular miRNA in this network and in tumorigenesis can be elucidated by first confirming the individual targets with which they interact. Algorithmic methods, based on predicted base pairing stability, can be used to identify potential protein targets [20], while experimental methods can be used to verify those targets [19]. This was the approach we used to investigate the role of let-7c-5p in clinical stage I ccRCC. One protein target that is predicted to interact with let-7c-5p is the insulin-like growth factor 1 receptor (IGF1R) [21,22,23,24,25]. IGF1R is a transmembrane receptor with tyrosine kinase activity when bound to insulin-like growth factor 1 (IGF-1) that shares high structural homology with the insulin receptor [26]. A study in head and neck squamous cell carcinoma has experimentally validated the interaction between IGF1R and let-7c-5p, connecting silenced expression of IGF1R to decreased proliferation, migration, and epithelial-mesenchymal transition (EMT) [27]. The expression levels of let-7c-5p have also shown a correlation with IGF1R through Western blotting in dental pulp-derived mesenchymal stem cells [28] and stem cells from the apical papilla [29]. IGF1R knockdown has been demonstrated to produce a similar impact in ccRCC cells. Aberrant IGF1R expression has been demonstrated to contribute to the malignant transformation of renal cells through its impact on cell proliferation, dedifferentiation, and apoptosis [26]. Additionally, siRNA-induced knockdown of IGF1R has been demonstrated to abrogate cell migration and invasion [30,31]. IGF1R expression levels within ccRCC have been shown to be prognostic with respect to patient outcomes. Patients with high IGF1R expression experience significantly poorer cancer-specific survival, with a 70% increased risk of death [32]. IGF1R staining correlates with Fuhrman grading in ccRCC tissues [33]. Furthermore, IGF1R overexpression corresponds to poor prognosis in other cancers, such as breast cancer, where it regulates EMT [34]. The objective of this study is to identify alterations in miRNA expression profiles of clinical T1 ccRCC that are characteristic of pathologic upstaging to pT3 at the time of surgery. In addition, we seek to better understand the role of let-7c-5p in ccRCC. Although this miRNA has been identified as a biomarker of aggressiveness in ccRCC and associated with the invasive phenotype in other tumor types, its specific mechanisms of action in ccRCC are not clearly delineated. Therefore, we aimed to further understand the role of let-7c-5p expression on the proliferation, migration, and invasion of ccRCC cells, along with its potential interaction with IGF1R. We investigated the link between let-7c-5p and IGF1R in order to improve our understanding of how each factor influences the progression of small ccRCC tumors and contributes to local invasion.
This study was performed according to the guidelines and regulations of the Lahey Hospital & Medical Center (LHMC) Institutional Review Board (IRB) protocol 2002-080 and the Boston Medical Center (BMC) IRB protocol H-37859. Although localized tumors as large as 7 cm can be classified as stage I, this study was limited to tumors ≤5 cm with the goal of characterizing the miRNA expression of small renal masses with or without an invasive phenotype. Twenty-four aged and ISUP grade-matched patients with cT1 renal masses who had undergone partial or radical nephrectomy for ccRCC at Lahey Hospital and Medical Center and Boston Medical Center were identified. Ten pT1 cases and ten pT3 cases were included in this study. Two patients were excluded due to the preoperative presence of metastasis, while two others were excluded due to synchronous tumors. A pathologist specializing in urologic malignancies (EJB & CSR) reviewed all samples to confirm clear cell histology, pathologic stage I or III disease, and assign an ISUP grade.
RNA was isolated from patient samples using the Qiagen FFPE AllPrep RNA isolation kit (Qiagen, Hilden, Germany) and was analyzed for quantity and purity (OD 260/280 ratio) using the Epoch spectrophotometer (BioTek, Winooski, VT, USA). Screening analysis was performed using Human miRNome panels I&II version 5 (Qiagen), analyzing 752 miRNAs via qRT-PCR according to the manufacturer’s protocol.
The human ccRCC cell lines 786-O (ATCC® CRL-1932™) and Caki-1 (ATCC® HTB-46™) (American Type Culture Collection, Manassas, VA, USA) were cultured under standard conditions (37 °C, 5% CO2). A VHL mutant primary epithelial clear cell adenocarcinoma with altered HIF and VEGF pathways [35], 786-O, was grown in RPMI 1640 (ATCC). Caki-1, a VHL wild type model line of metastatic ccRCC with high VEGF production [35], was maintained in McCoy’s 5A media (ATCC). All media were supplemented with 10% fetal bovine serum, penicillin/streptomycin, and L-glutamine.
A target search for predicted protein targets was conducted for let-7c-5p using a combination (accessed on 5 May 2021) of TargetScan (http://www.targetscan.org/) [21], miRmap (https://mirmap.ezlab.org/) [22], miRDB (http://mirdb.org/) [23,24], and miRcode (http://www.mircode.org/) [25] analyses. We chose a target gene based on predicted interaction across all search engines with let-7c-5p.
A hemocytometer (American Optical Corporation, Buffalo, NY, USA) was used to count cells and seed 786-O and Caki-1 into CELLSTAR six-well dishes (Greiner, Kremsmunster, Austria) at a density of 2 × 104 cells/mL. siPORT NeoFX transfection reagent (Invitrogen, Carlsbad, CA, USA) was used to deliver 2 μM pre-let-7c-5p (Cat. #AM17100, Assay ID PM10235, Ambion, Austin, TX, USA) or 2 μM pre-miR-Precursor Negative Control #1 (Cat. #AM17110, Ambion) at a final concentration of 20 nM to cells according to the manufacturer’s protocol. The Qiagen miRNeasy mini kit (Qiagen) was used to harvest RNA from transfected cells, and the Epoch spectrophotometer (BioTek, Winooski, VT, USA) was used to evaluate quality and purity (OD 260/280 ratio). Overexpression of let-7c-5p compared to control, RNU43, was conducted using Taqman miRNA qRT-PCR assays according to the manufacturer’s protocol (assay ID: 000379 and 001095, respectively, Applied Biosystems, Foster City, CA, USA). The comparative CT method was used to normalize let-7c-5p expression to RNU43 [36].
In vitro migration and invasion assays were conducted using 24 well plates containing Transwell (8 μm pores; Corning Costar Corp., Cambridge, MA, USA) membrane filter inserts 48 h after transfection. For the migration assay, chambers were placed in wells filled with serum-free medium supplemented with fibronectin (10 μg/mL). Cells (1 × 105 cells/mL) were added to the upper surface of the membrane and allowed to migrate through the membrane for 20 h at 37 °C. The invasion assay followed a similar protocol, albeit with the upper surface of the insert coated with 1:80 Matrigel (Becton-Dickinson, Franklin Lakes, NJ, USA) in serum free media. The inserts were allowed to rest in individual wells containing 10% RPMI for 786-O and 20% McCoy’s for Caki-1 cells. Caki-1 (5 × 105 cells/mL) or 786-O (1 × 105 cells/mL) cells were added to each Transwell chamber and allowed to invade through the membrane for 30 h at 37 °C. For each assay, cells that traversed the membrane were fixed in 10% w/v neutral-buffered formalin (Simport, Quebec, QC, Canada) and stained with DAPI (Invitrogen) (1:500 dilution in PBS, 1% Triton X-100). Cells were visualized using the EVOS FL microscope (Advanced Microscopy Group, Bothwell, WA, USA) and cell counts were recorded from three unique frames captured from each transwell, and a total of three transwells per condition.
Cells were transfected with either pre-let-7c-5p or pre-miR-Precursor Negative Control #1 (NC) and were seeded into 35 mm dishes (Corning) at a density of 2 × 104 cells/mL, in duplicate wells for each. RealTime-Glo™ MT Cell Viability Assay (Promega, Madison, WI, USA) reagents were added at a 1:2000 dilution 48 h after transfection. One hour after the addition of the assay reagents, luminescence readings were obtained on the GloMax® 20/20 Luminometer (Promega). This process was repeated 72 h after transfection.
Cell lysates were prepared from dishes displaying 70–80% confluency 48 h after transfection. Cells were lysed in 100 μL/well of boiled 1× SDS-Laemmli (250 mM Tris-HCl, 4% SDS, 10% glycerol, 0.003% bromophenol blue) by scraping the dishes manually, followed by shearing with a 24 gauge needle. A BCA assay (Pierce, Waltham, MA, USA) determined the total protein concentration of each sample. Paired lysates from NC and let-7c-5p cells were standardized for total protein concentration and volume and loaded in a unique well of a 12–230 kDa separation module on the Simple Western Jess system (ProteinSimple, Santa Clara, CA, USA). IGF1R primary antibody (NBP1-77679, Novus Biologicals, Littleton, CO, USA) at a concentration of 1:25 diluted in Antibody diluent 2 (ProteinSimple) and rabbit secondary antibody (042-206, ProteinSimple, used as provided) were added according to the manufacturer’s protocol. IGF1R values were normalized against total protein concentration and determined in the same capillary using RePlex and Total Protein Detection reagents (ProteinSimple).
Due to the large size of the 3′UTR of human IGF1R (7115 bp), analysis was performed on three distinct constructs comprising its full length (see Supplemental Materials). The constructs were inserted into vectors linked to the Firefly luciferase gene in addition to the Renilla luciferase gene for normalization (Genecopoeia, Rockville, MD, USA). Each fragment of the IGF1R 3′UTR contained a unique theoretical interaction site between let-7c-5p and the IGF1R gene. Additional mutant constructs were created through site-directed mutagenesis to assess each potential binding site. Caki-1 cells were grown in CELLSTAR 24-well dishes (Greiner) and, upon displaying 70–80% confluency, cell transfection was performed. At this time, the medium was replaced with Opti-MEM™ Reduced Serum Medium (31985062, Invitrogen), and transfected with a vector (0.2 μg) along with either pre-let-7c-5p or pre-miR-Precursor Negative Control #1 (30 nM), delivered with Endofectin (Genecopoeia). Cells were lysed according to the manufacturer’s protocol 24 h following transfection. Firefly and Renilla luminescence was measured using the Luc-Pair™ Duo-Luciferase Assay kit (Genecopoeia) in the GloMax® 20/20 Luminometer (Promega).
Categorical clinical variables were analyzed using Fisher’s exact tests, and the continuous clinical variables were analyzed using a two-tailed Welch’s t-test (SPSS v26). For the miRNA screening analysis, expression levels were normalized to the global mean of each sample for all miRNAs with a Cq < 35 for all samples. Expression levels were quantified as x = 2^(−ΔCt) and log2 transformed. Two-sided Welch’s t-tests were performed on the transformed values. Significant differences in miRNA expression levels were determined after adjusting for multiple comparisons using the Benjamini and Hochberg False Discovery Rate (FDR) method: presented as q-values. Hierarchical clustering was performed using GENE-E (Broad) with mean-centered values. For cell culture assays, a two-tailed Welch’s t-test was conducted to determine whether a statistically significant difference in traits exists for cells undergoing treatment compared to negative control. Corresponding plots depict the mean relative response rate for treated cells relative to negative control. Error bars represent the standard error of the mean. A p-value < 0.05 was considered statistically significant. All assays were performed with at least four separate experiments for each cell line. KEGG (Kyoto Encyclopedia of Genes and Genomes) analysis [37] was performed using DIANA-mirPath v. 3.0 [38]. The analysis was performed using the differentially expressed miRNAs of the let-7 family identified with a p-value < 0.05 and a q-value (FDR) < 0.25. A heat map was generated with the significant clusters resulting from pathways union analysis using FDR correction with a modified Fisher’s Exact Test, p < 0.05. Predicted gene targets with significant enrichment for these miRNAs were based on the miRNA interactions derived from microT-CDS with a threshold setting of 0.8 [39].
Clinical and pathological characteristics related to the tumor samples used in this study are detailed in Table 1. Patients were matched for age and tumor grade and, accordingly, no significant differences were noted for these selected criteria. In addition, no significant difference between the cohorts was observed for tumor size: mean of 3.7 cm for pT1 and 4.1 cm for pT3 tumors.
Of the 752 miRNAs analyzed, 108 miRNAs were detected for all samples (Ct < 35). Twenty miRNAs were found to be differentially expressed between the pT1 and pT3 small renal masses (p < 0.05 and FDR(q) < 0.25, Table 2). Hierarchical clustering analysis resulted in two distinct clusters differentiating these groups (Figure 1A). Of these 20 miRNAs, seven were upregulated and 13 were downregulated in pT3, compared to the pT1 lesions. Ten of these miRNA had a FDR < 0.05: let-7a-5p, -7c-5p, -7e-5p, miR-24-3p, 25-3p, -26a-5p, -26b-5p, -27a-3p, -93-5p, and -148b-3p. Specifically, let-7c-5p was downregulated 1.6 fold (p = 0.0023, FDR = 0.0393, Figure 1B) in this cohort of pT3 tumors, and let-7c-5p discriminated between these groups with an AUC of 0.880 (95%CI 0.721-1.0, Figure 1C). Five of these dysregulated miRNA are members of the let-7 family: let-7a-5p, -7b-5p, -7c-5p, -7d-5p, and -7e-5p. The KEGG analysis of these miRNAs identified nine predicted pathways (Figure 1D). Many of these pathways are seemingly related to an invasive phenotype. The most significant pathway identified was “ECM receptor interaction” (Supplemental Materials). In addition, 698 predicted genes were identified as being potential targets of at least two members of the let-7 family, and all five were predicted to target IGF1R (Supplemental Materials).
Let-7c-5p expression was rescued in 786-O and Caki-1 cell lines with pre-let-7c-5p and compared with a scrambled sequence as a negative control (Supplemental Materials). Forty-eight hours post-transfection, cells were tested in the Promega Cell Viability Assay to assess proliferation. While transfection with let-7c-5p did not significantly alter the proliferation of 786-O and Caki-1 cells after 48 h, at 72 h, each cell line displayed significantly diminished proliferation when transfected with pre-let-7c-5p as opposed to negative control (Figure 2A). Cell migration and invasion assays were conducted to explore the role of let-7c-5p in the EMT phenotype. Transfection of 786-O and Caki-1 cells with pre-let-7c-5p compared to negative control significantly reduced both cell migration and invasion in each cell line (Figure 2B,C). The net effect of increasing let-7c-5p levels in these cells resulted in phenotypic changes consistent with reduced metastatic potential.
IGF1R was identified as a direct target of let-7c-5p by each of the prediction algorithms used in this study [21,22,23,24,25], and our analysis focused on the three sites consistently identified by all of these prediction algorithms. These sites are referred to as 99, 2619, and 6661, based on their location in the 3′UTR of IGF1R (see Figure 3A and Supplemental Material). To confirm the direct interaction between let-7c-5p and the 3′UTR of IGF1R, a dual luciferase reporter assay was conducted in Caki-1 cells (Figure 3A). Unique vectors were constructed, each comprising ~2400 bp of the 3′UTR of IGF1R (see Supplemental Material for details), with each site present on a distinctive construct. Relative luciferase activity was calculated for each group through the Renilla/Firefly luminescence ratio. A significant reduction in relative luciferase activity was observed following co-transfection with let-7c-5p and the 2619 plasmid compared to the negative control and the 2619 mutant plasmid. Conversely, no significant difference was detected between cells co-transfected with let-7c-5p and the mutant 2619 plasmid compared to the negative control. When let-7c-5p was transfected with either the wild-type (WT) 99 or 6661 vector, and their mutant versions, no significant change was noted for WT compared to negative control or mutant. Western blot analysis, using the Jess Simple Western System, was conducted to further assess the impact of let-7c-5p on IGF1R expression levels in ccRCC cells. Transfection of 786-O and Caki-1 cells with pre-let-7c-5p as opposed to a negative control pre-miRNA construct significantly decreased IGF1R protein expression (Figure 3B).
Recent studies have focused on the prognostic ability of miRNA expression signatures to predict patient outcomes and their influence on carcinogenesis. miRNA dysregulation drives altered protein expression, namely through tumor suppressors and oncogenes. A previous study from our laboratory investigated pT1 tumors, identifying a number of miRNA as dysregulated in tumors that progressed to metastatic disease [6]. The present study further investigated the role of miRNA expression in the metastatic phenotype of small ccRCC tumors. Comparing cT1 tumors classified as pT1 with those pathologically upstaged to pT3a at the time of surgery elucidated a number of dysregulated miRNA that may play a role in conferring invasive properties within the tumor. One miRNA that has been downregulated in both studies, within aggressive pT1 tumors as well as cT1 tumors that have been upstaged to pT3a at the time of surgery, is let-7c-5p. The present study further demonstrated the role of let-7c-5p expression in the metastatic phenotype in ccRCC cells. Caki-1 and 786-O cells transfected with pre-let-7c-5p displayed a markedly reduced rate of proliferation, migration, and invasion compared to cells transfected with a negative control construct. We also confirm the interaction between let-7c-5p and IGF1R, an oncogene transmembrane receptor that displays tyrosine kinase activity when bound to IGF-1 [26], in ccRCC cells. Through luciferase and western blot analysis, we verify that let-7c-5p decreases IGF1R expression by binding to its 3′UTR. We hypothesize that reduced expression of let-7c-5p contributes to the overexpression of IGF1R in ccRCC cells, which contributes to tumor progression and metastasis. Interestingly, four other members of the let-7 family were identified as being dysregulated in this study (let -7a, -b, -d, and -e). These miRNAs share the same seed sequence (GAUGGAG) as let-7c-5p that is complimentary to the 3′UTR of IGF1R. Conceivably, one or more of these other family members may also be targeting IGF1R in ccRCC and therefore may cooperate in an additive manner to enhance regulation of IGF1R—as well as potentially other oncogenes. KEGG analysis of these miRNAs predicted their involvement in pathways related to the invasive phenotype, such as PI3K-AKT signaling, Wnt signaling, and the most significant process in our analysis—ECM receptor interaction. PI3K-AKT and Wnt signaling pathways have been well studied in oncogenesis [40], and they have been implicated specifically in the invasive phenotype of ccRCC [41,42]. An examination of the ECM receptor interaction pathway (Supplemental Materials) identified proteins directly of interest to this study. Various collagens and integrins, most notably, are members of this pathway that are likely targeted by multiple members of the let-7 family dysregulated in this study (Supplemental Material). Integrins are cell surface molecules that are involved in cell–cell and cell extracellular matrix adhesion and play important roles in cell proliferation, migration, and signaling [43]. Cross-talk between integrins and the IGF1R signaling pathway has been well documented [44,45]. Additionally, intergrin expression levels have been correlated with patient survival and metastasis, which suggests that some integrins may play a vital role in cancer progression. Specific to renal cell carcinoma, Breuksch et al. reported that the integrin α5 expression correlated with distant metastases within five years after tumor nephrectomy and reduced survival [46]. Collagens and other associated ECM components also have the potential to influence IGF1R signaling [47]. Collagen levels have been correlated with cancer invasion and lymph node metastasis [48,49,50,51] and have been linked to poor prognosis in several cancers, including ccRCC [52,53]. The concept that ECM ligands bind to integrins and IGF1 binds to IGF1R has been proposed [54]. The let-7 family as well as some of its targets, including integrins and collagen, have all been associated with invasive and metastatic properties and linked to poor prognosis in ccRCC. Collectively, this evidence supports the role of the let-7 family as agents of tumor suppression in small renal masses. Thus, it is interesting to speculate on the therapeutic potential of inhibiting ccRCC tumors by systemic delivery of let-7 mimics, as demonstrated in a mouse model of lung cancer [55]. A limitation of this study is that we only investigated the binding of let-7c-5p in this work. It was the let family member that was most consistently dysregulated across this study and our previous investigation of aggressive small renal masses [6], and it is arguably the most noted let-7 family member reported as being dysregulated in ccRCC. Studies on let-7c-5p have demonstrated the miRNA to be downregulated in ccRCC as well as other cancer types. Peng et al. found let-7c-5p to be downregulated in ccRCC tissue compared to adjacent normal, and correlated the downregulation with increased chemoresistance to the drug 5-fluorouracil [11]. Additionally, let-7c-5p expression has been shown to be significantly decreased in early metastatic ccRCC tumors compared with non-metastatic tumors [9]. Low expression of let-7c-5p has been correlated with disease state in a variety of other cancers, such as prostate [13,14], acute promyelocytic leukemia [15], hepatocellular [16], non-small cell lung [17], and endometrial [18]. Specifically, in non-small cell lung cancer, Zhao et al. detected a significant association between let-7c-5p downregulation, venous invasion, and metastasis [17], which is consistent with our results within small ccRCC tumors. Conversely, while low let-7c-5p expression is frequently correlated with poor patient outcomes in human cancers, there is some evidence of let-7c-5p acting as an oncogene. High expression levels have been found in aggressive cancers with poorer patient prognosis in high grade serous ovarian carcinoma [56] and oral tongue squamous cell carcinoma [57]. While let-7c-5p undoubtedly targets the 3′UTR of many protein transcripts, we chose to focus on IGF1R because of its predicted interaction with let-7c-5p and its association with poorer prognosis for ccRCC patients. The link between these two factors has previously been established in head and neck squamous cell carcinoma via a luciferase reporter assay [27] but, to our knowledge, has not been investigated in ccRCC. IGF1R is a transmembrane receptor that exerts its influence on cell behavior through the activation of various downstream effectors such as the PI3K-AKT network, mitogen activated protein kinase (MAPK) pathway, and mTOR [58]. Expression levels of IGF1R correlate with Fuhrman grading in ccRCC tissues [33], and high IGF1R expression confers a significant reduction in cancer specific survival [32]. High expression of IGF1R has also been shown to induce clathrin-dependent endocytosis in which IGF1R is translocated to the nucleus, which is associated with adverse prognosis [59]. Heightened expression of IGF1R has also been identified within ccRCC biopsy compared to normal kidney, and inactivation of the von Hippel–Lindau tumor suppressor gene has been directly linked to this upregulation [60]. In ccRCC cells, siRNA-induced knockdown of IGF1R leads to diminished migratory and invasive abilities, along with a reduction in proliferation [30,31]. Additionally, Yuen et al. demonstrated the effects of IGF1R depletion on chemoresistance, as ccRCC cells became desensitized to 5-fluorouracil and etoposide [58]. Interestingly, both let-7c depletion and IGF1R overexpression have been shown to increase chemoresistance to 5-fluorouracil. These findings together suggest a potential therapeutic benefit from abrogating the dysregulation of these two factors. While we chose to focus on the role of let-7c-5p in the invasive capacity of small ccRCC tumors, other miRNA that have been implicated both in the present study as well as prior work from our laboratory on pT1 ccRCC. miR-25-3p, miR-93-5p, and miR-484 have all been identified as upregulated in aggressive ccRCC tumors, while miR-26a-5p, 26b-5p, miR-23b-3p, miR-27a-3p, miR-148b-3p, let-7e-5p, and let-7b-5p expression levels are reduced. Evaluation of the influence of these miRNA on ccRCC cell behavior and the protein targets that they regulate may improve our understanding of why these small tumors progress to metastatic disease.
In this study, we identify a subset of miRNA that is dysregulated in small (<5 cm) cT1 ccRCC that were later upstaged to pT3 at the time of surgery due to the presence of invasion. Members of the let-7c family were downregulated in these tumors, as well as in pT1 tumors that would later progress to metastatic disease. Dysregulation of these miRNAs facilitates the metastatic phenotype and portends a poor prognosis. The reduced expression of the let-7 family likely contributes to the upregulation of IGF1R in these tumors, and we demonstrate that let-7c-5p modulates proliferation, migration, invasion, and IGF1R expression in ccRCC cells. | true | true | true |
|
PMC9598763 | Hongyan Li,Fang Liu,Hanzhe Kuang,Hua Teng,Siyi Chen,Sijing Zeng,Qimin Zhou,Zhaokai Li,Desheng Liang,Zhuo Li,Lingqian Wu | WDR73 Depletion Destabilizes PIP4K2C Activity and Impairs Focal Adhesion Formation in Galloway–Mowat Syndrome | 25-09-2022 | WDR73,PIP4K2C,PIP2,focal adhesion | Simple Summary Galloway–Mowat syndrome is a rare genetic disease, classically characterized by a combination of various neurological symptoms and nephrotic syndrome. WDR73 is the pathogenic gene responsible for Galloway–Mowat syndrome. However, the pathological and molecular mechanisms of Galloway–Mowat syndrome, especially nephrotic syndrome caused by WDR73 deficiency, remains unknown. In this study, we knocked out the WDR73 in human embryonic kidney 293 cells to observe the morphological characteristics of the cells and elucidate the functions of WDR73. Additionally, we used a combination of proteomics, transcriptomics, and biochemical assays to identify the regulated targets of WDR73. We aimed to discover directly interacting molecules and the regulatory pathway of WDR73 and to illustrate the molecular mechanism between the WDR73 pathway and nephrotic disease in Galloway–Mowat syndrome. From the molecular mechanism we found in vitro, we draw a hypothesis that the damage to focal adhesion of podocytes caused by WDR73 defect is the key issue of kidney dysfunction. Finally, we verified the hypothesis in a podocyte-specific conditional knockout Wdr73 mouse model. Abstract (1) Background: Galloway–Mowat syndrome (GAMOS) is a rare genetic disease, classically characterized by a combination of various neurological symptoms and nephrotic syndrome. WDR73 is the pathogenic gene responsible for GAMOS1. However, the pathological and molecular mechanisms of GAMOS1, especially nephrotic syndrome caused by WDR73 deficiency, remain unknown. (2) Methods and Results: In this study, we first observed remarkable cellular morphological changes including impaired cell adhesion, decreased pseudopodia, and G2/M phase arrest in WDR73 knockout (KO) HEK 293 cells. The differentially expressed genes in WDR73 KO cells were enriched in the focal adhesion (FA) pathway. Additionally, PIP4K2C, a phospholipid kinase also involved in the FA pathway, was subsequently validated to interact with WDR73 via protein microarray and GST pulldown. WDR73 regulates PIP4K2C protein stability through the autophagy–lysosomal pathway. The stability of PIP4K2C was significantly disrupted by WDR73 KO, leading to a remarkable reduction in PIP2 and thus weakening the FA formation. In addition, we found that podocyte-specific conditional knockout (Wdr73 CKO) mice showed high levels of albuminuria and podocyte foot process injury in the ADR-induced model. FA formation was impaired in primary podocytes derived from Wdr73 CKO mice. (3) Conclusions: Since FA has been well known for its critical roles in maintaining podocyte structures and function, our study indicated that nephrotic syndrome in GAMOS1 is associated with disruption of FA caused by WDR73 deficiency. | WDR73 Depletion Destabilizes PIP4K2C Activity and Impairs Focal Adhesion Formation in Galloway–Mowat Syndrome
Galloway–Mowat syndrome is a rare genetic disease, classically characterized by a combination of various neurological symptoms and nephrotic syndrome. WDR73 is the pathogenic gene responsible for Galloway–Mowat syndrome. However, the pathological and molecular mechanisms of Galloway–Mowat syndrome, especially nephrotic syndrome caused by WDR73 deficiency, remains unknown. In this study, we knocked out the WDR73 in human embryonic kidney 293 cells to observe the morphological characteristics of the cells and elucidate the functions of WDR73. Additionally, we used a combination of proteomics, transcriptomics, and biochemical assays to identify the regulated targets of WDR73. We aimed to discover directly interacting molecules and the regulatory pathway of WDR73 and to illustrate the molecular mechanism between the WDR73 pathway and nephrotic disease in Galloway–Mowat syndrome. From the molecular mechanism we found in vitro, we draw a hypothesis that the damage to focal adhesion of podocytes caused by WDR73 defect is the key issue of kidney dysfunction. Finally, we verified the hypothesis in a podocyte-specific conditional knockout Wdr73 mouse model.
(1) Background: Galloway–Mowat syndrome (GAMOS) is a rare genetic disease, classically characterized by a combination of various neurological symptoms and nephrotic syndrome. WDR73 is the pathogenic gene responsible for GAMOS1. However, the pathological and molecular mechanisms of GAMOS1, especially nephrotic syndrome caused by WDR73 deficiency, remain unknown. (2) Methods and Results: In this study, we first observed remarkable cellular morphological changes including impaired cell adhesion, decreased pseudopodia, and G2/M phase arrest in WDR73 knockout (KO) HEK 293 cells. The differentially expressed genes in WDR73 KO cells were enriched in the focal adhesion (FA) pathway. Additionally, PIP4K2C, a phospholipid kinase also involved in the FA pathway, was subsequently validated to interact with WDR73 via protein microarray and GST pulldown. WDR73 regulates PIP4K2C protein stability through the autophagy–lysosomal pathway. The stability of PIP4K2C was significantly disrupted by WDR73 KO, leading to a remarkable reduction in PIP2 and thus weakening the FA formation. In addition, we found that podocyte-specific conditional knockout (Wdr73 CKO) mice showed high levels of albuminuria and podocyte foot process injury in the ADR-induced model. FA formation was impaired in primary podocytes derived from Wdr73 CKO mice. (3) Conclusions: Since FA has been well known for its critical roles in maintaining podocyte structures and function, our study indicated that nephrotic syndrome in GAMOS1 is associated with disruption of FA caused by WDR73 deficiency.
Galloway–Mowat syndrome (GAMOS; OMIM 251300) is a rare recessive genetic disease characterized by neurodevelopmental defects and progressive renal glomerulopathy. Neurological involvement of GAMOS contains microcephaly, developmental delay, intellectual disability, and other variable neural symptoms [1]. Renal manifestations range from isolated proteinuria to early nephrotic syndrome (NS), which may rapidly progress to end-stage renal disease. The prognosis of GAMOS is poor, and the majority of affected children died before six years of age due to therapy-resistant renal failure [2,3]. To date, mutations in 11 different genes, such as WDR73, KEOPS complex genes, YRDC, and NUP107, have been reported to cause GAMOS [4,5,6,7]. Although WDR73 is the first gene to be identified in GAMOS, the molecular mechanisms underlying the pathophysiology of GAMOS upon WDR73 function remain obscure. WDR73 encodes a WD40-repeat-containing protein that can be expressed in a variety of cells. In the fetal kidney, WDR73 exists in immature podocytes from the S-shaped body stage to the capillary-loop stage. In mature glomeruli, WDR73 has a punctate distribution at the periphery of the glomerular tuft and is present in the cell bodies of mature podocytes. WDR73 is localized in the cytoplasm during interphase and accumulates at spindle poles and microtubule asters during mitosis; thereby, it may play a role in cellular architecture and cell cycle [4]. It is reported that WDR73 could interact with α- and β-tubulin in fibroblasts [8]. Loss of WDR73 also has been demonstrated to disrupt the integrator complex, perturbing the transcription of cell cycle regulatory proteins [9]. Despite these observations providing some insights into the function of WDR73, the regulatory pathway of WDR73, the precise contributions of WDR73 to cell physiological function, and the mechanisms underlying GAMOS, especially nephrotic syndrome caused by WDR73 deficiency, are poorly understood. In this study, we knocked out the WDR73 in human embryonic kidney (HEK) 293 cells to observe the morphological characteristics of the cells and elucidate the functions of WDR73. Additionally, we used a combination of proteomics, transcriptomics, and biochemical assays to identify the regulated targets of WDR73. We aimed to discover directly interacting molecules and the regulatory pathway of WDR73 and to illustrate the molecular mechanism between the WDR73 pathway and nephrotic disease in Galloway–Mowat syndrome. From the molecular mechanism we found in vitro, we draw a hypothesis that the damage to focal adhesion of podocytes caused by WDR73 defect is the key issue of kidney dysfunction. Finally, we verified the hypothesis in a Wdr73 CKO mouse model.
HEK 293 cells were cultured in DMEM medium supplemented with 10% FBS and 1% penicillin–streptomycin (Thermo Fisher Scientific, Waltham, MA, USA) at 37 °C under 5% CO2. Transfections were performed using Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, USA), following the manufacturer’s instructions. The following reagents were used in this study: cycloheximide (CHX), MG132, and chloroquine (Selleck, Houston, TX, USA).
HEK293-WDR73 KO cells were generated using the CRISPR/Cas9 system. Briefly, plasmids carrying sgRNA targeting the WDR73 region encompassing exon 6 were introduced into HEK293 cells using Lipofectamine 3000. Transfected HEK293 colonies were selected using puromycin (Solarbio, Beijing, China). The puromycin-selected colonies were seeded into 96-well plates (1 cell per well) and expanded. Furthermore, WDR73 with frameshift deletions in the six exons was confirmed by Sanger sequencing. Western blotting was used to examine WDR73 protein expression in each clone. WDR73 KO cells were transfected with plasmids carrying WT WDR73 using Lipofectamine 3000 and selected using hygromycin. Cell lines carrying the stably integrated plasmid were expanded from a single colony. The colony with an approximate protein expression level was selected.
Plasmids were constructed as follows: Plasmids expressing WT WDR73 were generated by cloning human WDR73 into pcDNA3.1-Hygro vector. For co-immunoprecipitation, a plasmid expressing 3×HA-WDR73 was generated by cloning human WDR73 into an in-house modified version of the pcDNA3.1(+) -3×HA vector. PCMV-GFP-hEPG5 was kindly provided by Professor Hong Zhang (Institute of Biophysics, Chinese Academy of Sciences, China). Plasmids expressing GFP were generated by cloning GFP into a PCMA vector. For the GST pulldown assay, PIP4K2C was PCR-amplified and cloned into the pGEX-6P-1 vector to produce GST-tag fused recombinant proteins, whereas WDR73 was PCR-amplified and cloned into the pET28A vector to produce His-tag fused recombinant proteins. Plasmids for WDR73 KO were constructed by inserting sgRNA sequences into the pSpCas9(BB)-2A-Puro (PX459) V2.0 plasmid. The sgRNA sequences specifically targeting WDR73 were as follows: F, CACCGACTTCGGAGCCTCGCCCCA; R, AAACTGGGGCGAGGCTCCGAAGTC. The antibodies used in this study are shown in Table S4.
The proliferative ability of the cells was measured using the CCK-8 assay (Vazyme, Nanjing, China) according to the manufacturer’s instructions. Briefly, cells were plated into a 96-well plate at a density of 2.0 × 103 cells/well and incubated at 37 °C for 0, 24, 48, 72, and 96 h. Twenty microliters of CCK-8 reagent was added to each well, and the cells were cultured for 2 h. All experiments were performed in triplicates. Absorbance was measured at 450 nm using a microplate reader (ELx800, BioTek, Winooski, VT, USA).
The cells were harvested after 72 h, and the cell suspension was then digested. Next, the cells were fixed using ethanol (75%) for 4 h at 4 °C, and the supernatant was subsequently discarded, followed by incubation with an RNA enzyme containing iodide (PI, Sigma-Aldrich, St. Louise, MO, USA). After the cells were washed three times with PBS, the cell cycle was detected using a Cytek Dxp Athena flow cytometer (Cytek, Biosciences, San Diego, CA, USA), and data analysis was conducted using the Modfit LT software. For the determination of apoptosis, the cells were stained with FITC-conjugated annexin V and PI according to the manufacturer’s instructions (Vazyme, Nanjing, China). Data were collected and analyzed using a Cytek Dxp Athena flow cytometer, and data analysis was performed using the FlowJo software 10.4 (BD, Ashland, USA). All experiments were performed in triplicates.
For all cell lines, 2.0 × 103 cells/well were seeded onto 96-well plates, and four regions per well were imaged every 2 h over a period of 48 h using the Incucyte Zoom (Essen BioScience, Ann Arbor, MI, USA). The cell areas were measured using the Incucyte Zoom software package to obtain quantitative data on the extent of cell spreading.
The coverslip-grown cells were fixed in 4% paraformaldehyde for 20 min at room temperature or in ice-cold MeOH for 5 min. The coverslips were permeabilized in 0.5% Triton X-100/PBS for 15 min and then blocked in 5% bovine serum albumin/PBS for 1 h at room temperature. Primary antibodies were applied in a dilution according to the instructions on staining buffer overnight at 4 °C. Secondary antibodies (Alexa Fluor secondary 488; cy3, Jackson, West Grove, PA, USA) were applied in a 1:200 dilution in staining buffer for 1 h at 37 °C in the dark. The nuclei were stained with DAPI. Actin filaments were labeled with FS488 phalloidin (Solarbio, Beijing, China). All images were captured using a confocal laser scanning microscope (Zeiss LSM 880 with airyscan, Zeiss, Berlin, Germany). The primary antibodies used were described previously.
The recombinant His-WDR73 fusion protein was purified by MerryBio Co., Ltd. (Huai’an, China). After quantification and qualification, the recombinant His-WDR73 fusion proteins were labeled with Cy3 (CyDye Protein LabellingCY3 MONO 5-PACK, GE) and used to probe the Arrayit HuProt v4.0 20 K Human Proteome Microarrays (CDI Laboratories, Baltimore, MD, USA). Briefly, Cy3-tagged His-WDR73 was incubated on the microarray overnight at 4 °C and washed. Finally, the microarray was scanned using a microarray scanner (CapitalBio, Beijing, China). Data were analyzed to generate a candidate list of WDR73-binding proteins; the signal-to-noise ratio (SNR) was defined as the ratio of the median foreground value to the median background value. A total of 336 candidates were selected based on a Z-score > 3 (Z-score = (SNR-mean)/SD). Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were conducted to reveal the unique biological significance and key pathways associated with WDR73 in the candidate list of WDR73-binding proteins (criteria: p-value < 0.05, significantly enriched).
Cells were harvested and lysed using RIPA lysis buffer (Beyotime, Haimen, China) containing 1 mM PMSF and 1× protease inhibitor cocktail (Sigma-Aldrich, St. Louise, MO, USA). Cell lysates were separated using SDS-PAGE and transferred to a PVDF membrane (Merck Millipore, Burlington, MA, USA). The membranes were blocked in Tris-buffered saline containing 0.01% Tween20 and 5% non-fat milk for 1 h and then incubated with specific primary antibodies. Following incubation with horseradish peroxidase (HRP)-linked secondary antibody (Jackson, West Grove, PA, USA) at room temperature for 1 h, detection was performed using an Immobilon Western Chemiluminescent HRP substrate kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions. The primary antibodies used were as previously described.
The plasmids encoding GST, GST-PIP4K2C, and His-WDR73 were transfected into E. coli. The fusion proteins were prepared as previously described. Approximately 100 µg of GST and GST-PIP4K2C fusion protein was immobilized in 50 µL of Mag-Beads GST Fusion Protein Purification (Sangon Biotech, Shanghai, China) and equilibrated before being incubated together at 4 °C for 60 min with a gentle rocking motion. Approximately 100 µg of His-WDR73 fusion protein was added to immobilized GST-PIP4K2C and GST after three washes with PBST. The two fusion proteins were incubated overnight at 4 °C with gentle agitation. The bound proteins were washed five times with PBS, boiled with loading buffer for 5 min, and analyzed by Western blotting, as described previously.
HEK293 cells were transfected with plasmids expressing 3HA-WDR73, GFP-Epg5, 3HA-WDR73, and GFP using Lipofectamine 3000 (Invitrogen Waltham, MA, USA). Cells were lysed using immunoprecipitation buffer (20 mM Tris (pH7.5), 150 mM NaCl, 1% Triton X-100, 1 mM PMSF, and 1× protease inhibitor cocktail (Sigma-Aldrich St. Louise, MO, USA) 48 h after transfection. For Co-IP, approximately 1000 μg protein was incubated at 4 °C overnight with 30 μL indicated monoclonal Anti-HA magnetic beads (Bimake, Houston, TX, USA). The precipitates were washed three times with immunoprecipitation buffer, boiled with loading buffer for 5 min, and analyzed by Western blotting, as described previously.
PIP2 was quantified using a PI(4,5)P2 mass ELISA kit (echelon, K-4500, Salt Lake City, UT, USA). Briefly, lipids were extracted from WDR73 KO and WT cells according to the manufacturer’s instructions. The PIP2 levels were measured according to the manufacturer’s instructions. The absorbance was measured at 450 nm using a microplate reader (ELx800, BioTek, Winooski, VT, USA). Finally, the quantity of PIP2 was calculated using a standard curve.
Total RNA was extracted from WDR73 KO and WT cells using TRIzol reagent (Invitrogen, Waltham, MA, USA). After RNA quantification and qualification, library preparation and transcriptome sequencing were performed using Novogene Bioinformatics Technology (Beijing, China). Differential expression analysis was performed using the DESeq2 R package. Finally, enrichment analysis of differentially expressed genes (DEGs), GO enrichment analysis, and KEGG pathway analysis were conducted using clusterProfiler (Bioconductor, Boston, MA, USA) to reveal the unique biological significance and key pathways associated with WDR73 in the DEGs (criteria: p-value < 0.05, significantly enriched). The Disease Ontology (DO) database describes the function of human genes and diseases, and we used clusterProfiler software to test the statistical enrichment of DEGs in the DO pathway (criteria: p < 0.05, significantly enriched).
Total RNA was extracted from cells using TRIzol reagent (Invitrogen, Waltham, MA, USA), and mRNA was reverse transcribed using a RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions. The real-time PCR assay was conducted using 2× SYBR Green Mix (Thermo Fisher Scientific, Waltham, MA, USA) on an ASA-9600 qRT-PCR System (Lanzhou Baiyun Gene Technology Co., Ltd, Lanzhou, China). Primers used for real-time PCR are listed in Tables S2, S3 and S5. Relative expression was calculated using the 2−ΔΔCt method.
Mice were housed in a specific-pathogen-free facility and kept in a 12 h day/night cycle with free access to chow and water. Genotyping and breeding of animals were performed according to standard procedures. Wdr73 general KO mice were generated by Shanghai Biomodel Organisms Center using CRISPR/Cas9. A guide RNA targeting the KO exon 6–exon 8 of the Wdr73 (ENSMUST00000026816.14) gene was designed. The guide RNA1 and guide RNA2 sequences are shown in Table S5. Male Wdr73 frameshift heterozygous mice and female Wdr73 frameshift heterozygous mice were mated to obtain homozygous mice. For genotype identification, the Wdr73 frameshift mutation was identified by PCR amplification using the primers PI, PII, PIII, and PIV; (Table S5). All mouse lines were maintained in the C57BL/6J background by regular backcrossing to the C57BL/6J line. Study protocols complied with all relevant ethical regulations and were approved by the IRB of Central South University (IRB: 2019-2-17). Wdr73 flox mice were generated by GemPharmatech Co., Ltd. (Nanjing, China), using CRISPR/Cas9 on a C57BL/6 background. These mice carry a cassette with LoxP sites flanking a region containing 154 bp coding sequence of Wdr73 exon 4–exon 5, the knocking out of which will result in disruption of protein function. For the generation of a podocyte-specific Wdr73 knockout mouse model, Wdr73flox/flox mice were crossed with Npsh2-Cre mice on a C57BL/6J background (GemPharmatech Co., Ltd, Nanjing, China). In the nephropathy experiments, male and female mice (aged 5–8 weeks), were treated with a single retroorbital injection of ADR (doxorubicin HCl; Macklin, shanghai, China) at the dose of 20 mg/kg for Nphs2-Cre, Wdr73flox/flox (Wdr73 CKO), and Wdr73flox/flox (WT) mice. For genotype identification, the Wdr73 flox was identified by PCR amplification using the primers Wdr73-5F/R and Wdr73-3F/R; Cre recombinase was identified by PCR amplification using the primers Nphs2-cre F/R (Table S5). General PCR amplification was performed to verify Wdr73 gene knockout efficiency using the primers Wdr73-5F and Wdr73-3R (Table S5). The age of animals used for the respective experiments is stated in the figures and/or figure legends (male and female animals showed similar phenotypes and were combined for the analysis).
Albumin and creatinine levels were quantified by measuring spot urine from Wdr73 CKO and WT mice at defined time points. Proteinuria was expressed as albumin–creatinine ratio. Assessment of urinary albumin was performed using a mouse-specific, albumin fluorescence-based kit (Fankewei, Shanghai, China). Measurement of creatinine was performed using an enzymatic creatinine kit (Fankewei, Shanghai, China).
First, 16 μL of collected spot urine was mixed with 4 μL 4 × LDS Sample Buffer (NP0007, Invitrogen, Waltham, MA, USA); the sample was heated at 70 °C for 10 min for optimal results and separated by SDS–PAGE. Gels were stained with Coomassie for 1 h and destained using standard methods.
Fresh kidneys underwent primary fixation with 2% glutaraldehyde in PBS. They were then postfixed in 1% osmium tetroxide for 1 h and dehydrated in 50%, 70%, 90%, 95%, or 100% ethanol and propylene oxide for 10 min each. Samples were further infiltrated with an epoxy resin mixture. Ultrathin sections were collected on copper grids, and sections were stained using 10% uranyl acetate in 50% methanol and modified Sato lead stain. A HITACHI HT7700 electron microscope was used for picture acquisition (Lab of Biomedical Electronic Microscopy Higher Research Center, Central South University, Changsha, China).
Fresh kidney immersion was fixed in 4% PFA in phosphate-buffered saline (PBS) for 24 h and subsequently dehydrated in 20% or 30% sucrose solution for 24 h each at 4 °C. Kidney cryosections of O.C.T.-embedded (20 μm thick) (Sakura TissueTek #4583, Sakura Finetek, Alphen aan den Rijn, the Netherlands) and kidney sections (3–4 μm thick) of paraffin-embedded (FFPE) tissue were generated using standard methods. FFPE sections were deparaffinized and rehydrated and then underwent heat-induced antigen treatment.
Glomeruli were isolated from male or female mice aged between 5 and 8 weeks old using previously described methods [10]. Mice are anesthetized with Avertin (250 mg/kg body weight) and perfused via the left heart ventricle with 20–40 mL PBS. Isolated kidneys were minced in ice PBS and digested with collagenase type II (2 mg/mL; C5138, Sigma-Aldrich St. Louise, MO, USA) at 37 °C for 30 min, and glomeruli were sequentially sieved with a 100 μm and 70 μm cell strainer in order and washed with PBS. The isolated glomeruli were seeded on collagen I (C8062, Solarbio, Beijing, China)-coated plates. The method yields primary podocytes with 90% purity, confirmed via staining with the podocyte-specific marker nephrin on day 24 after isolation. Primary mouse podocytes were cultured in RPMI medium (Gibco, Waltham, MA, USA) supplemented with 10% FBS and 1% penicillin–streptomycin (Thermo Fisher Scientific Waltham, MA, USA) at 37 °C under 5% CO2, and they were fed with fresh medium every 2–3 days.
Podocyte proliferation was measured using the EdU assay kit (C0071S, Beyotime, Haimen, China). The podocytes were fixed with 0.5% buffered paraformaldehyde and then incubated with the EdU detection solution for 30 min in the dark. Hoechst (C0071S, Beyotime, Haimen, China) was applied to the podocytes, which were incubated in the dark for 10 min. Finally, cell proliferation was analyzed by fluorescence microscopy (Zeiss LSM 880 with airyscan, Zeiss, Berlin, Germany).
To investigate the role of WDR73 in the maintenance of normal cellular physiological functions, we knocked out WDR73 in HEK293 cells using CRISPR/Cas9 with a single-guide RNA (sgRNA) that targets exon 6 of WDR73 (Figure S1). Subsequently, knock-in with WT WDR73 was performed to further confirm the specific effect caused by WDR73 KO and eliminate off-target effects or other factors in the treatment process (Figure 1A). The effect of WDR73 on the growth and proliferation of cells was then investigated, and the results of the CCK-8 assay revealed that WDR73 KO cells showed a relatively slow growth rate compared with WT cells, which could be partly rescued by WDR73 knock-in (Figure 1B). Additionally, flow cytometry analysis indicated that WDR73 KO led to a remarkable increase in G2/M cell percentages. This blocking of G2/M could also be rescued by WDR73 knock-in (Figure 1C). However, apoptosis of KO cells was not significantly different from that of WT cells (Figure S2). These data indicated that WDR73 is essential for cell growth and proliferation.
Due to the increase in the G2/M phase in WDR73 KO cells, we tried to define the potential roles of WDR73 in cell division and the features underlying delayed mitotic progression. Spindle morphology was examined by immunofluorescence with α-tubulin and γ-tubulin, revealing that the percentage of abnormal bipolar spindles of WDR73 KO cells significantly increased (Figure 2A and Figure S3). Some WDR73 KO cells exhibited shorter spindles than the WT cells (Figure 2B). Metaphase spindles in cultured adherent cells commonly adopt a planar orientation parallel to the surface of the culture dish. We observed that WT cells assembled a bipolar spindle oriented parallel to the substratum; in contrast, WDR73 KO cells had a bipolar spindle oriented unparallel to the substratum, such that spindle poles were detected in a different confocal z-axis. Our measurements of the spindle angle from planar orientation showed that WDR73 KO cells exhibited significantly larger spindle angles than WT cells (Figure 2C). These results indicate that WDR73 plays an important role in establishing spindle orientation and polarity axis for cell division. Moreover, we observed that WDR73 KO caused changes in cell shape and size and inhibited cell adhesion and spreading on the dishes. Four hours after plating, a portion of WT and rescued cells adhered to the bottom of the plate, whereas the WDR73 KO cells maintained a round shape, floating in the medium. Another four hours later, most of the WT and rescued cells completed the adhesion process. Some of them even started cell division. Simultaneously, only a few parts of the KO cells began to adhere. WDR73 KO cells also exhibited delayed cell spreading and thus had a smaller projected area than WT cells (Figure 2D,E). At 48 h after cell plating, microfilaments were labeled with 488-phalloidin, showing that WDR73 KO cells displayed different cell morphologies and a lower degree of cell spreading compared to WT cells. Pseudopodia formation in WDR73 KO cells was significantly reduced by disturbed F-actin assembly (Figure 2F and Figure S4). These results suggested that WDR73 KO causes delayed cell adhesion and spreading, implying a role of WDR73 in cell polarization and migration.
RNA sequencing (RNA-seq) was used to provide insights into the transcriptome of the WDR73 KO cells. Three WDR73 KO samples and three WT samples were analyzed. The classical Bayesian algorithm was applied to identify DEGs, and volcano plots were used to visualize the variation in DEGs between WDR73 KO cells and WT cells. In total, 1042 mRNAs were differentially expressed in WDR73 KO cells, including 707 upregulated and 335 downregulated mRNAs (Figure S5). Gene enrichment analysis was used to assess the functional associations of DEGs derived from WDR73 KO cells. KEGG analysis results showed that DEGs were mostly enriched in FA, the MAPK signaling pathway, and ECM–receptor interaction (Figure 3A). Additionally, GO enrichment analysis results showed that the DEGs were mostly enriched in the extracellular matrix (ECM, belonging to cellular component) (Figure 3B). These analyses suggested that many of the DEGs were crucial for FA and ECM (Figure 3C). The two pathways had nine overlapping DEGs, which were all validated to be consistent with the RNA-seq data using qRT-PCR (Figure S6). Since FA complexes provide the main sites of cell adhesion to the ECM and are associated with the actin cytoskeleton, the interaction between FA and ECM plays a critical role in cell spreading, adhesion, and pseudopodia formation [11,12]. RNA-seq analysis revealed that the DEGs related to cell adhesion and ECM disturbed pseudopodia formation, which is supported by our previous findings that WDR73 KO inhibited cell adhesion and spreading. Disease Ontology (DO) analysis results showed that DEGs could be significantly enriched in kidney disease (Figure 3D). This pathway and FA had significantly overlapping DEGs (p = 6.10 × 10−7) (Figure 3E), indicating that DEGs related to the FA pathway may be involved in kidney disease.
Although both the cell morphological alteration and DEG pattern suggest that WDR73 KO may disrupt FA and ECM-related pathways, the regulatory mechanism needs to be explored. We used the HuProt (Žilina, Slovakia) human protein microarray to identify potential WDR73-interacting proteins. A protein microarray composed of 21,000 purified human proteins as N-terminal GST fusions was incubated with purified human WDR73 protein. In total, 336 proteins were identified as potential WDR73-interacting proteins. To gain insight into the functional roles of WDR73, KEGG pathway analysis was conducted on 336 proteins to reveal their unique biological significance and key pathways. Potential WDR73-interacting proteins can be enriched in the regulation of the actin cytoskeleton pathway (Figure 4A), which plays an important role in cell adhesion, spreading, division, cell polarity, and spindle orientation [13,14]. Of the proteins involved in the regulation of the actin cytoskeleton pathway (Table S1), PIP4K2C generated the greatest signal-to-noise ratio (SNR 21.33) by protein microarray (Figure 4B). PIP4K2C and PIP4K2A (SNR 9.47) are two kinases generating PIP2, and previous immunoprecipitation experiments showed that PIP4K2C could interact with active PIP4K2A in vitro and can heterodimerize with PIP4K2A [15,16]. PIP2 plays a key role in restructuring the actin cytoskeleton [17,18]. Hence, we confirmed the potential direct interaction of WDR73 with PIP4K2C using the GST pulldown assay. Bacterially expressed GST-tagged PIP4K2C, but not GST alone, pulling His-tagged WDR73 down (Figure 4C) indicated that these two proteins can interact directly with each other. Moreover, Western blotting showed that WDR73 KO led to a significant reduction in the PIP4K2C and PIP4K2A protein levels, which can be rescued partly by WDR73 knock-in (Figure 4D). Thus, we assessed whether the decrease in PIP4K2C and PIP4K2A levels caused by WDR73 KO actually reduced intracellular PIP2. As shown in Figure 4E, WDR73 KO significantly reduced intracellular PIP2 production. Evidence suggests that PIP2 plays important roles in regulating actin reorganization and FA assembly [18,19,20]. To investigate the formation of FAs in cells, cells were stained with an anti-paxillin antibody, which showed that FAs were widely distributed and aggregated into spots in WT cells. However, the intensity of paxillin foci decreased remarkably in the pseudopodia of WDR73 KO cells, which might indicate a decrease in FA formation (Figure 4F and Figure S7). The decreased FA was not due to the reduced expression of FA core proteins such as talin, paxillin, vinculin, and FAK (Figure S8). As several kinases, including type I (PIP5K) and type II (PIP4K) kinase isoforms (A, B, and C), are involved in generation of PIP2, we also investigated the mRNA transcription level of these kinases in different organs of the mice. Results revealed the highest transcription of PIP4K2C among the tested kinases in the kidney (Figure S9), implying that it has a specific physiological function in the kidney and may participate in the progression of kidney disease in GAMOS1.
Since there was a significant reduction in PIP4K2C and PIP4K2A protein levels due to the WDR73 depletion, the mechanism underlying the decrease in such proteins induced by WDR73 KO was systematically explored. First, quantitative real-time PCR analysis showed that WDR73 depletion did not significantly affect PIP4K2C mRNA levels (Figure 5A); therefore, downregulation of PIP4K2C by WDR73 KO should occur at the post-transcriptional stage. Subsequently, we assessed the PIP4K2C stability in WDR73 KO cells after treatment with cycloheximide (Figure 5B), a protein synthesis inhibitor. The results showed that the PIP4K2C protein degraded faster in WDR73 KO cells, suggesting that the stability of PIP4K2C was impaired by WDR73 depletion. The ubiquitin–proteasome system (UPS) and autophagy–lysosomal pathway (ALP) are the two major protein degradation systems in eukaryotic cells [21]. When WDR73 KO and WT cells were treated with the UPS inhibitor MG132 or the ALP inhibitor chloroquine, downregulation of PIP4K2C by WDR73 depletion was restored by chloroquine, but not by MG132 (Figure 5C), indicating that WDR73 might affect the autophagy–lysosome-dependent degradation of PIP4K2C. To determine if the WDR73 KO would induce an abnormal increase in autophagy, the abundance of microtubule-associated protein 1 light chain 3 II (LC3II) and autophagy substrate protein p62 was detected. Western blotting showed that the rate of LC3-I to LC3-II conversion increased, and p62 levels were dramatically reduced in WDR73 KO cells (Figure 5D). By re-querying the data from the previous human proteome microarray, a protein named ectopic P-granules autophagy protein 5 (EPG5) with a high SNR (SNR: 33.64) was also identified to potentially interact with WDR73 (Figure 5E). The Co-IP assay also confirmed the interaction between WDR73 and EPG5 (Figure 5F). These results revealed that WDR73 KO could enhance autophagic flux and reduce the stability of PIP4P2C and PIP4K2A proteins through EPG5.
To further confirm the role of Wdr73 in vivo, we first generated WDR73 systemic KO mice. However, no Wdr73-/- embryos and pups were detected at any given period (Figure S10), suggesting that loss of WDR73 is embryonically lethal in mice, and WDR73 might play an important role in early development. Podocytes are essential for the maintenance of the glomerular filtration barrier. Therefore, we generated a podocyte-specific conditional knockout mouse of Wdr73 (Wdr73 CKO) by crossing Wdr73flox/flox with Nphs2-Cre to study the role of WDR73 in podocytes (Figure 6A,B). The Wdr73 CKO mice did not show any growth or histologic lesion abnormalities, and they exhibited no increase in proteinuria levels up to the age of 20 weeks (Figure 6C,D and Figure S11A,B). Then, we applied Adriamycin (ADR) to induce glomerular injury both for the WT and Wdr73 CKO strains to confirm the role of WDR73 in maintaining podocyte function (Figure 6E). We found that loss of WDR73 significantly increased the susceptibility of podocytes toward injury by ADR, which was characterized by the detection of high levels of albuminuria and podocyte injury in Wdr73 CKO mice at 5 weeks after ADR injection. ADR-treated Wdr73 CKO mice exhibited increased albuminuria levels in both male and female Wdr73 CKO mice compared with the same treated WT mice of the same sex (Figure 6F,G). Podocyte injury was evidenced by glomerular basement membrane (GBM) thickening and podocyte foot process (FP) broadening and effacement by transmission electron microscopy (TEM) analysis (Figure 6H). However, kidneys of ADR-treated Wdr73 CKO mice did not show obvious histologic alterations in the glomerular or tubular compartment (Figure 6I and Figure S11C). Together, these observations suggested that WDR73 deficiency causes impaired podocyte FP formation, resulting in an impaired filtration barrier in the kidney.
To further dissect the effect of WDR73 deletion in podocytes, we established a primary podocyte culture system, and cells were confirmed to be positive for nephrin expression before the assay (Figure S12). First, we labeled the cells with the proliferation marker EdU, and the results showed that the proliferation rate of primary podocytes isolated from ADR-treated Wdr73 CKO mice was reduced compared with WT cells (Figure 7A,B). In addition, ADR-treated Wdr73 CKO mice primary podocytes showed marked differences in FAs when compared to WT podocytes. FAs were widely distributed and assembled into spots in the WT podocytes, whereas in the Wdr73 CKO mouse primary podocytes, the FA structure was disrupted, and the number and size of speckle-like adhesion spots were reduced (Figure 7C,D).
In this study, we revealed that WDR73 exerts an essential function in the regulation of focal adhesion and the actin cytoskeleton and provides a novel molecular mechanism in reduced renal cell proliferation, adhesion, and spreading. By human protein microarray assay and GST pulldown, we confirmed that WDR73 can directly interact with PIP4K2C, which usually heterodimerizes with PIP4K2A and catalyzes the phosphorylation of phosphatidylinositol 5-phosphate (PI5P) to generate PIP2. WDR73 depletion led to a significant reduction in the PIP4K2C and PIP4K2A proteins, consequently inducing a dramatic decrease in intracellular PIP2. PIP2 plays a key role in FA assembly and actin polymerization [18]. FA complexes provide the main sites for cell adhesion to the ECM and are associated with the actin cytoskeleton, which is required for cell adhesion and spindle orientation in cell division [14,22,23]. Actually, FA and ECM alterations in podocytes have been widely reported to be one of the main cellular bases for proteinuria. FA is a key signal and structural hub of foot processes that regulate the actin cytoskeleton, through which podocytes could firmly bind to the GBM, which is a dense matrix of ECM components. Once the FA dynamic or assembly is disrupted, podocytes gradually detach from the GBM, leading to the irreversible progression of kidney disease [24,25,26]. We demonstrated that there were significant changes in cell morphology of WDR73 KO cells, with significantly reduced cell spreading area and less pseudopod formation. The adhesion ability also decreased dramatically. We observed that the number of speckle-like adhesion spots was significantly reduced and the shape was ambiguous in WDR73 KO cells. Considering the RNA-seq results, the enrichment of DEGs in FA and ECM-related pathways together, these results suggest that FA and actin filament assembly were disrupted by the absence of WDR73, leading to pseudopod formation, cell adhesion, and spreading abnormality, which may be the primary cellular pathological basis for the albuminuria and other renal abnormalities in GAMOS1. Although this point has not been mentioned in other studies on GAMOS, we confirmed that the FA formation and distribution were impaired in primary podocytes derived from ADR-injected Wdr73 CKO mice. In fact, despite the resistant C57BL/6 background, Wdr73 CKO animals with only a single dose of ADR manifested increased albuminuria levels, GBM thickening, and podocyte effacement, all suggesting podocyte injury. Taken together, these findings strongly indicate that WDR73 plays key physiologic and cell biologic roles in focal adhesion. PIP4K2C is ubiquitously expressed at different levels in tissues, with primary expression in the kidney, brain, heart, and testes [15]. However, the role of PIP4K2C in kidney development and disease remains unclear. PIP4K2C can heterodimerize with PIP4K2A and has been suggested to serve as a chaperone for the more active isoforms, and its presence likely affects the distribution and enzyme activity of the 2A and 2B isoforms [16,27]. We confirmed that WDR73 can directly interact with PIP4K2C, and we found that the mRNA expression of PIP4K2C was higher in the mouse kidney than in other organs. WDR73 KO led to a significant reduction in the PIP4K2C and PIP4K2A proteins, resulting in the reduction in intracellular PIP2 and disruption of FA and actin filament assembly. These results may provide new clues regarding the role of interaction between PIP4K2C and WDR73 in the kidney. Considering the reduction in PIP4K2C induced by WDR73 depletion, the stability of PIP4K2C was assessed. Our results revealed the unstable PIP4K2C should be eliminated mainly through enhanced autophagy, but not the ubiquitin–proteasome system. An autophagy protein EPG5 was subsequently proved to also interact with WDR73. EPG5 is a Rab7 effector that determines the fusion specificity of autophagosomes with late endosomes/lysosomes and is a key autophagy regulatory protein [28]. We intended to further confirm these molecular interactions with WDR73 in vivo using mice for further verification.
In summary, our study suggested that WDR73 can interact with PIP4K2C to form a complex with a function in FA assembly and regulates PIP4K2C protein stability through the autophagy–lysosomal pathway. This study would provide a new perspective on the molecular mechanism of WDR73 in primary albuminuria and other kidney abnormalities of GAMOS1. | true | true | true |
|
PMC9598820 | Xiao Yang,Wangjie Jiang,Xiangxu Kong,Xiao Zhou,Deming Zhu,Lianbao Kong | Genistein Restricts the Epithelial Mesenchymal Transformation (EMT) and Stemness of Hepatocellular Carcinoma via Upregulating miR-1275 to Inhibit the EIF5A2/PI3K/Akt Pathway | 22-09-2022 | hepatocellular carcinoma,miR-1275,epithelial mesenchymal transformation,stemness,EIF5A2 | Simple Summary Genistein is a natural phytoestrogen with various antitumor effects. Our study focused on exploring the mechanisms of microRNAs and genistein to inhibit the epithelial mesenchymal transformation (EMT) and stemness of hepatocellular carcinoma (HCC). We found that miR-1275 was more highly expressed in HCC cells treated with genistein compared with the control. Then, we performed series functional experiments to explore the relationship between genistein and miR-1275 in HCC. The inhibition of genistein on HCC cells was enhanced by the increase in treatment time and dose, and miR-1275 can be raised by genistein. The overall survival and recurrence-free survival of HCC patients with low expressed miR-1275 were lower than those of those with high expression levels. The experimental results exhibited that genistein and miR-1275 can both significantly suppress the proliferation, migration, invasion, metastasis, EMT and stemness of HCC. Moreover, the inhibition can be further enhanced with the co-existence of miR-1275 mimic and genistein. Finally, we demonstrated that miR-1275 can inhibit the EMT and stemness of HCC via inhibiting the EIF5A2/PI3K/Akt pathway. Our findings proved that genistein can inhibit the EIF5A2/PI3K/Akt pathway by upregulating miR-1275 so as to attenuate the EMT and stemness of HCC cells to restrict their progression and metastasis. Abstract Purpose: Genistein is a natural phytoestrogen with various antitumor effects. In recent years, some microRNAs (miRNA) in cancer cells have been reported to be regulated by genistein. Our study focused on exploring the mechanisms of miRNA upregulation to inhibit the epithelial mesenchymal transformation (EMT) and stemness of hepatocellular carcinoma (HCC). Patients and Methods: MiR-1275 was discovered by the transcriptome sequencing of miRNA expression profiles in HepG2 cells treated with genistein or DMSO as a control. Then, we performed series functional experiments in vitro and vivo to explore the relationship between genistein and miR-1275 in HCC. The target gene (Eukaryotic initiation factor 5A2, EIF5A2) of miR-1275 was predicted by databases and finally determined by a dual luciferase reporter assay. The downstream signaling pathway of EIF5A2 was assessed by bioinformatics analysis and Western blot. Results: the inhibition of genistein on the viability of HCC cells was enhanced by the increase in treatment time and dose, but it had no obvious inhibitory effect on normal hepatocytes (QSG-7701). Through qRT-PCR and transcriptome sequencing, we discovered that miR-1275 was lowly expressed in HCC, and it can be raised by genistein. The overall survival (OS) and recurrence-free survival (RFS) of HCC patients with lowly expressed miR-1275 were lower than those of those with high expression levels. In vitro and vivo experiments exhibited that genistein and the overexpression of miR-1275 can both significantly suppress the proliferation, migration, invasion, metastasis, EMT and stemness of HCC. Moreover, the inhibition can be further enhanced when miR-1275 mimic and genistein exist together. Finally, we demonstrated that miR-1275 can inhibit the epithelial mesenchymal transformation (EMT) and stemness of HCC via inhibiting the EIF5A2/PI3K/Akt pathway. Conclusion: Our findings proved that genistein can inhibit the EIF5A2/PI3K/Akt pathway by upregulating miR-1275 so as to attenuate the EMT and stemness of HCC cells to restrict their progression and metastasis. | Genistein Restricts the Epithelial Mesenchymal Transformation (EMT) and Stemness of Hepatocellular Carcinoma via Upregulating miR-1275 to Inhibit the EIF5A2/PI3K/Akt Pathway
Genistein is a natural phytoestrogen with various antitumor effects. Our study focused on exploring the mechanisms of microRNAs and genistein to inhibit the epithelial mesenchymal transformation (EMT) and stemness of hepatocellular carcinoma (HCC). We found that miR-1275 was more highly expressed in HCC cells treated with genistein compared with the control. Then, we performed series functional experiments to explore the relationship between genistein and miR-1275 in HCC. The inhibition of genistein on HCC cells was enhanced by the increase in treatment time and dose, and miR-1275 can be raised by genistein. The overall survival and recurrence-free survival of HCC patients with low expressed miR-1275 were lower than those of those with high expression levels. The experimental results exhibited that genistein and miR-1275 can both significantly suppress the proliferation, migration, invasion, metastasis, EMT and stemness of HCC. Moreover, the inhibition can be further enhanced with the co-existence of miR-1275 mimic and genistein. Finally, we demonstrated that miR-1275 can inhibit the EMT and stemness of HCC via inhibiting the EIF5A2/PI3K/Akt pathway. Our findings proved that genistein can inhibit the EIF5A2/PI3K/Akt pathway by upregulating miR-1275 so as to attenuate the EMT and stemness of HCC cells to restrict their progression and metastasis.
Purpose: Genistein is a natural phytoestrogen with various antitumor effects. In recent years, some microRNAs (miRNA) in cancer cells have been reported to be regulated by genistein. Our study focused on exploring the mechanisms of miRNA upregulation to inhibit the epithelial mesenchymal transformation (EMT) and stemness of hepatocellular carcinoma (HCC). Patients and Methods: MiR-1275 was discovered by the transcriptome sequencing of miRNA expression profiles in HepG2 cells treated with genistein or DMSO as a control. Then, we performed series functional experiments in vitro and vivo to explore the relationship between genistein and miR-1275 in HCC. The target gene (Eukaryotic initiation factor 5A2, EIF5A2) of miR-1275 was predicted by databases and finally determined by a dual luciferase reporter assay. The downstream signaling pathway of EIF5A2 was assessed by bioinformatics analysis and Western blot. Results: the inhibition of genistein on the viability of HCC cells was enhanced by the increase in treatment time and dose, but it had no obvious inhibitory effect on normal hepatocytes (QSG-7701). Through qRT-PCR and transcriptome sequencing, we discovered that miR-1275 was lowly expressed in HCC, and it can be raised by genistein. The overall survival (OS) and recurrence-free survival (RFS) of HCC patients with lowly expressed miR-1275 were lower than those of those with high expression levels. In vitro and vivo experiments exhibited that genistein and the overexpression of miR-1275 can both significantly suppress the proliferation, migration, invasion, metastasis, EMT and stemness of HCC. Moreover, the inhibition can be further enhanced when miR-1275 mimic and genistein exist together. Finally, we demonstrated that miR-1275 can inhibit the epithelial mesenchymal transformation (EMT) and stemness of HCC via inhibiting the EIF5A2/PI3K/Akt pathway. Conclusion: Our findings proved that genistein can inhibit the EIF5A2/PI3K/Akt pathway by upregulating miR-1275 so as to attenuate the EMT and stemness of HCC cells to restrict their progression and metastasis.
Hepatocellular carcinoma (HCC) ranks second among all cancers that can cause cancer-related death [1]. The poor prognosis of HCC is attributed to multiple pathogenic factors, the high recurrence and the metastasis rate [2]. Although great progress has been made in the research of surgical strategies and targeted therapeutic drugs, the coping strategies for the recurrence and metastasis of HCC are still limited [3]. Epithelial mesenchymal transformation (EMT) and stemness are the key factors leading to a high recurrence and metastasis rate of HCC. EMT refers to the process of transforming tumor cells with epithelial characteristics into cells with mesenchymal characteristics. In this process, tumor cells lose the polarity and the ability of adhesion to the basement membrane and acquire stronger characteristics of mobility and metastasis [4]. The stemness of tumor cells is defined as the ability of self-renewal, multi-lineage differentiation potential and tumor initiation, which is essential for tumor recurrence and metastasis. Furthermore, the stemness is plastic and can be affected by diversified factors, such as hypoxia and EMT [3,5,6]. It has been reported that the activation of classical signaling pathways responsible for EMT such as Akt, TGF-β, C-MYC and Notch can also improve the stemness of tumor cells [7]. The enhancement of EMT and stemness usually leads to drug resistance and worse clinical outcomes, so it is imperative to find novel adjuvant drugs for the treatment of HCC. Genistein, a natural compound extracted from soybean, is the member of the soybean isoflavone family with the most extensive biological activities and antitumor effects [8]. Previous reports have shown that genistein has the functions of protecting the cardiovascular system, regulating blood glucose, delaying neurological decline and antitumor [9,10,11,12]. In addition, recent studies on genistein have confirmed that it possesses the ability to inhibit the proliferation, invasion and metastasis of a variety of cancers, including HCC, renal cancer and breast cancer [8,13,14]. At the same time, several studies have shown that genistein can enhance the efficacy of sorafenib and reverse the resistance of HCC to sorafenib [13,15]. These reports also indicated that genistein has potential value as a new adjuvant therapy for HCC. However, the specific mechanism of genistein inhibiting tumor progression and metastasis has not been fully elucidated, especially in HCC. In our study, we speculated whether genistein can prevent the progression and metastasis of HCC by inhibiting the EMT and stemness. MicroRNAs (miRNAs) also play an important role in the regulation of the EMT and stemness of HCC, so we focused our research on the relationship between genistein and miRNAs in HCC. MiRNAs are a group of small single stranded noncoding RNAs that are 20–24 bp in length and silence the translation products of the targeted mRNAs to adjust cell life activities. MiRNAs participate in the post transcriptional regulation of target genes, mainly by binding to the 3′UTR region of target mRNAs [16,17]. In these studies of HCC, many miRNAs have been proven to be involved in the regulation of tumor proliferation, invasion and metastasis [16,17,18]. For example, Qiongying Hu et al. [19] confirmed that miR-101 can suppress HCC progression by decreasing the expression level of VEGF. MiRNAs may become novel breakthroughs or therapeutic targets for seeking new therapeutic methods for HCC. Previous studies mainly focused on the direct effects of genistein on HCC. However, whether genistein can restrain the progression and metastasis of HCC via regulating miRNAs has not been determined. Therefore, in this study, we screened the upregulated miRNAs in HCC cells treated with genistein by sequencing and further explored whether these miRNAs play an antitumor role by inhibiting EMT and stemness. Our results revealed a new pathway for genistein to inhibit HCC progression and metastasis and also provide a theoretical basis for exploring new treatments for HCC.
Seventy pairs of HCC samples and their paracancerous tissues were obtained from the Department of Hepatobiliary Center, the First Affiliated Hospital of Nanjing Medical University. All HCC patients underwent surgery and were diagnosed by pathology. This study was authorized in advance by the Department of Institutional Review Board (IRB), the First Affiliated Hospital of Nanjing Medical University, and all patients and their guardians signed Informed consent forms.
Normal hepatocytes (QSG-7701), HCC cell lines (HCC-LM3, MHCC-97H, Hep-3B, Hep-G2 and Huh-7) and HEK293T cells were acquired from the Institute of Cell Biology, Chinese Academy of Sciences (Shanghai, China). All the above cell lines were cultured in a DMEM medium (Gibco BRL, Gaithersburg, MD, USA) supplemented with high glucose, 10% fetal bovine serum (FBS) (TransGen Biotech, Beijing, China) and 1% Penicillin-Streptomycin (TransGen Biotech, Beijing, China). All these cells were cultured in a cell incubator containing 5% CO2 at 37 °C before experiments.
Genistein (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in dimethylsulfoxide (DMSO) before treating cells, and the volume fraction of DMSO was less than 0.1% in the culture medium. According to the results of cell viability experiments, the HCC cells were treated with genistein (10.81 µg/mL) or DMSO (the control group) 3 days in advance of the experiments.
Inhibitor, inhibitor-NC, Mimic and Mimic-NC of miR-1275 (RiboBio, Guangzhou, China) were transfected into HCC cells by Lipofectamine 3000 (Invitrogen, Waltham, MA USA). After 3 days, these HCC cells were used for subsequent functional experiments. In rescue experiments, the overexpression plasmid of Eukaryotic initiation factor 5A2 (EIF5A2) (CoreusBiotech, Nanjing, China) was also transfected into HCC cells by Lipofectamine 3000. The scramble plasmid of EIF5A2 (EIF5A2-NC) (CoreusBiotech, Nanjing, China) was used as the control. These cells transfected with mimics or plasmids were prepared for subsequent in vitro experiments.
The lentivirus vectors encoding homo ov-miR-1275 (CoreusBiotech, Nanjing, China) were the constructed overexpression miR-1275, and the empty vectors (ov-vector) were used as the control group. Then, the HCC cells were transfected with lentivirus for 72 h. Then, the transfection efficiency was observed by an immunofluorescence microscope. Finally, the HCC cells were screened with puromycin (5 µg/mL) for 3 weeks until stable cell clones were established. These cells transfected with lentivirus were used for subsequent in vivo experiments.
HCC cells and normal hepatocytes (QSG-7701) were seeded into 96-well cell culture plates (3000 cells/well) and then separately treated with genistein at the concentrations of 0, 0.1, 0.5, 1, 5, 10, 50 and 100 µg/mL for 3 days. The medium was replaced with a fresh culture medium containing 10% CCK-8 solution (Vazyme, Nanjing, China) at 24 h, 48 h and 72 h, respectively, and incubated with cells at 37 °C for 2 h. The absorbance of each well was measured at 450 nm by an automatic microplate reader (ELX808, BioTek, Norcross, GA, USA) for further half maximal inhibitory concentration (IC50) analysis.
According to the results of the cell viability test, the miRNA expression profile in Hep-G2 cells treated with genistein (10.81 µg/mL) or DMSO (control) for 3 days was analyzed by high throughput transcriptome sequencing (LC-Bio, Hangzhou, China). Three independent samples were prepared for each group in advance.
The total RNA isolated from samples was performed following the instructions of the RNA Extraction Kit (Vazyme, Nanjing, China). The reverse transcription synthesis of cDNA was operated under the guidance of the HiScript II Q RT SuperMix kit (Vazyme, Nanjing, China). The expression levels of miR-1275 and EIF5A2 were measured by qRT-PCR, and this assay was conducted by the 7900HT Fast Real-Time PCR System (ABI, Waltham, MA, USA) and ChamQ Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China). U6 was used as the internal reference of miR-1275, and GAPDH was selected to normalize EIF5A2. The sequences of all primers (RiboBio, Guangzhou, China) are provided in the Supplementary Materials (Table S1).
The HCC cells in each group were seeded into 96-well culture plates (2000 cells/well) and cultured for 4 days. These cells were supplemented with a fresh culture medium containing 10% CCK-8 solution (Vazyme, Nanjing, China) at 0 h, 24 h, 48 h, 72 h and 96 h, respectively, and incubated at 37 °C for 2 h. Then, the absorbance of each well was detected at 450 nm by an automatic microplate reader (ELX808, BioTek, Norcross, GA, USA).
The HCC cells of each group in 96-well culture plates (104/well) were labeled with EDU and incubated at 37 °C for 2 h. Then, these cells were fixed with 5% paraformaldehyde and stained according to the steps in the instructions of the BeyoClickTM EDU-488 Cell Proliferation Kit (Beyotime, Shanghai, China). Finally, these cells were stained with DAPI for 30 min and supplemented with fresh phosphate buffered saline (PBS) (Gibco BRL, Gaithersburg, MD, USA). The number of HCC cells stained with EDU was observed under the fluorescence microscope and calculated by Image J software.
The HCC cells in each group were added into 6-well culture plates (Corning, Corning, NY, USA). When the density of HCC cells exceeded 95%, the tip of a 200 μL pipette gun was applied to make a scratch in the center of each well. The healing stages of scratches at 0 h, 24 h, 48 h and 72 h were observed and photographed under an inverted microscope (ZEISS, Oberkochen, Germany), and the proportion of the healing area was measured by Image J software v1.8.0 (National Institutes of Health, Bethesda, MD, USA).
The HCC cells (2 × 104) and 250 uL serum-free medium were added into the upper part of each transwell chamber (8 um pore size; Corning, Corning, CA, USA). At the same time, 600 uL medium with 20% FBS was supplemented into the lower part of each chamber. After 24 h, these chambers were fixed with 5% paraformaldehyde for 30 min and then stained with crystal violet solution for 30 min. After these chambers were washed, the excess cells in the upper part of the chambers were wiped off with a cotton swab, and the migrated cells were photographed under an upright microscope. Finally, the number of migrated cells in each chamber was measured by Image J software.
This test was performed in 24-well ultra-low adsorption culture plates with a U-shaped bottom (Corning, USA) and applied to evaluate the stemness of HCC cells. Before this test, the specific culture medium was prepared with insulin (4 ng/mL), B27, EGF (80 ng/mL), bFGF (10 ng/mL) and DMEM/F12 medium (Gibco, Waltham, MA, USA) [20]. Then, 500 μL special medium containing HCC cells (1000) was added to each well, and the medium was replaced every 2 days. After 1 week, the number of spheroids in each well was photographed and counted by an inverted microscope (ZEISS, Germany).
This test was conducted to assess the invasive ability of HCC cells and operated in 96-well culture plates (Corning, USA). The HCC cell spheroids were centrifuged and resuspended in DMEM medium for use, and the pH value of type I collagen (Corning, USA) was adjusted to 7–7.5. Then, the HCC cell spheroid suspension and type I collagen solution were mixed in equal proportion and added to 96-well culture plates. Then, 150 μL DMEM culture medium was added into each well after the plates were incubated at 37 °C for 30 min. After 2 days, the invasion area of each spheroid was observed by an inverted microscope (ZEISS, Germany) and counted by Image J software.
This assay was performed to evaluate the effect of genistein and miR-1275 on the DNA damage and apoptosis of HCC cells. The Comet Assay was conducted on the basis of the instructions of the Comet Assay kit (KeyGEN BioTECH, Nanjing, China). The HCC cells immobilized on glass slides containing three layers of agarose gel were electrophoresed in an alkaline solution for 20 min and washed three times with a neutralizing solution. Then, these cells were stained with DAPI for 30 min and photographed under a fluorescent microscope (Leica, Wilmington, NC, USA). The Olive tail moment (the degree of DNA damage) was measured by CASP software.
This assay was performed to explore the effects of genistein and miR-1275 on the cell cycle and proliferation of HCC cells. The HCC cells (105/well) in each group were seeded into 6-well culture plates (Corning, USA) and treated with genistein (10.81 µg/mL) or transfected with miR-1275 mimic. After 3 days, these HCC cells were trypsinized, fixed with absolute ethanol and stored at −20 °C. Then, HCC cells were washed and stained with PropidiumIodide (PI) following the guidance of the instructions of the Cell Cycle Staining Kit (MULTISCIENCES, Hangzhou, China). The distribution of the HCC cell cycle was counted by the FACSCantoTM II flow cytometer (BD, Franklin Lakes, NJ, USA), and the results were analyzed by ModFit LT software 3.1 (Verity Software House, Mountain View, CA, USA).
The primary antibodies involved in the Western blot (WB) are as follows: anti-EIF5A2 (1:2000), anti-PI3K (1:1000), anti-Akt (1:1000), anti-p-Akt (1:1000), anti-SOX2 (1:1000), anti-BMI1 (1:1000) and anti-OCT4 (1:1000) were used (Abcam, Cambridge, UK); anti-E-cadherin (E-cad) (1:1000), anti-N-cadherin (N-cad) (1:1000), anti-Vimentin (Vim) (1:1000) and anti-β-actin were used (Proteintech, Wuhan, China); the secondary antibody was HRP-IgG (1:3000) (Proteintech, Wuhan, China).
There were 20 6-week-old male BALB/c nude mice purchased from the Animal Experiment Center of Nanjing Medical University (Nanjing, China), and they were randomly assigned to 4 groups (ov-vector, ov-miR-1275, ov-vector + Genistein and ov-miR-1275 + Genistein, n = 5). These nude mice were subcutaneously injected with HCC-LM3 cells (106 cells/100μL, 100 μL/mouse). Before injection, HCC-LM3 cells were transfected with the lentivirus of ov-vector or ov-miR-1275 in advance. In the genistein treatment group, each nude mouse was intraperitoneally injected with 100 μL genistein (10.81 µg/mL) solution every 3 days. After 35 days, these subcutaneous HCC tumors obtained from the sacrificed mice were fixed with formaldehyde and embedded in paraffin. These specimens were sectioned and stained with haematoxylin & eosin (H&E) and immunohistochemistry (IHC) for EIF5A2, Ki-67, PI3K, p-Akt, SOX2 and Vimentin. This animal experiment was approved by the Animal Ethics Committee of Nanjing Medical University (Nanjing, China).
As described above, a total of 20 6-week-old male BALB/c nude mice were randomly divided into 4 groups. These nude mice were injected with HCC-LM3 cells via the caudal vein, and the administration method of genistein was the same as that described above. After 50 days, these mice were euthanized, and their lung tissues were removed and photographed. These lung tissues were sectioned and stained with H&E. The numbers of metastatic nodules in lung tissues were counted under an orthographic microscope. This animal experiment was also approved by the Animal Ethics Committee of Nanjing Medical University (Nanjing, China).
The plasmids (labeled with firefly luciferase) containing wild-type or mutant fragments of EIF5A2 3′-UTR and miR-1275 mimic or NC duplex (labeled with Renilla luciferase) were co-transfected to HEK293T cells for 2 days. Then HEK293T cells were lysed, and the luciferase activity was measured under the guidance of the instructions of the Bio-Lite Luciferase Assay System (Vazyme, Nanjing, China) by a fluorometer.
CancerMIRNome (http://bioinfo.jialab-ucr.org/CancerMIRNome/, accessed on 9 September 2021) was used to analyze the expression level and prognosis of miR-1275 in The Cancer Genome Atlas (TCGA) database. Timer (cistrome.shinyapps.io/timer) was used to analyze the expression level of EIF5A2 in the TCGA database. The Kaplan–Meier (K-M) Plotter (http://kmplot.com/analysis/, accessed on 1 October 2010) was applied to analyze the effect of the EIF5A2 expression level on the survival of HCC patients in the TCGA-LIHC database. StarBase (https://starbase.sysu.edu.cn/, accessed on 1 December 2013) was adopted to predict the correlation between EIF5A2 and tumor stemness markers. Four databases (TargetScan, miRWalk, miRDB and miRPathDB) were used to predict the target genes of miR-1275. EMTome (http://www.emtome.org/, accessed on 10 December 2020) was applied to predict the downstream signaling pathways of EIF5A2, and DIANA-miRPath v3.0 (http://www.microrna.gr/miRPathv3/, accessed on 1 July 2015) was used to predict the downstream signaling pathways of miRNAs. The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway and Gene Set Variation Analysis (GSVA) enrichment analyses of miR-1275 and EIF5A2 were performed by R software version 4.1.0 (Robert Gentleman, Auckland, New Zealand).
All assays were repeated at least three times, and the mean and curve graphs showed the mean and standard deviation. The data were analyzed using the Chi-squared test or Student’s t-test by Graphpad prism software v8.0.2 (GraphPad Software, San Diego, CA, USA). Pearson’s correlation test was applied to analyze the correlation between EIF5A2 and miR-1275. Statistical significance was defined as p < 0.05 (*), p < 0.01 (**) and p < 0.001 (***).
First, we found that genistein had strong inhibitory effects on the viability of different HCC cell lines but had no significant effect on QSG-7701 cells. Interestingly, the inhibition was enhanced by the increase in time and the concentration of genistein. Hep-G2 cells were the most sensitive to genistein, and the IC50 of genistein in Hep-G2 cells at 72 h was 11.74 µg/mL. On the contrary, the IC50 of genistein in HCC-LM3 cells at 72 h was the highest (19.57 µg/mL) (Figure 1A–C). Therefore, we selected Hep-G2 cells treated with genistein for miRNA transcriptome sequencing, and HCC-LM3 cells were also used for subsequent experimental verification. The heatmap of sequencing results was displayed as follows (Figure 1D). The differentially expressed miRNAs between HCC samples (n = 372) and normal liver tissues (n = 50) in the TCGA-LIHC database were selected depending on the threshold (p-value < 0.05 and |Log2(FoldChange)| > 1) (Figure 1E). By drawing a Venn diagram, we discovered that only miR-1275 was positively regulated by genistein and was lowly expressed in HCC tissues in the TCGA-LIHC database (Figure 1F). Further pan-cancer analysis also exhibited that miR-1275 is mainly lowly expressed in a variety of tumor samples, especially in HCC tissues (Figure 1G,H). The K-M survival analysis of miR-1275 in HCC patients in the TCGA database showed that, although the LogRank test showed no significant difference (p-value = 0.22), the overall survival (OS) of HCC patients with a high expressed miR-1275 was higher than that of those with a low expressed miR-1275 (Figure 1I). The receiver operating characteristic (ROC) curve of the miR-1275 expression level in HCC patients in the TCGA-LIHC database was also provided, and the area under curve (AUC) was 0.73 (Figure 1J). Next, Hep-G2 and HCC-LM3 cells were treated with genistein or DMSO. As expected, the miR-1275 levels were more highly expressed in the genistein treatment group than they were in the DMSO control group (Figure 1K). However, instead, the expression level of miR-1275 changed inconspicuously in QSG-7701 cells after the treatment with genistein (Figure S1B). Similarly, in our single center, miR-1275 was shown to be more lowly expressed in HCC tissues than that in paracancerous tissues (n = 70) by qRT-PCR (Figure 1L). HCC patients were also grouped, and the clinical data were counted according to the expression level of miR-1275. The results were listed as a baseline table (Table 1). It showed that a low expressed miR-1275 was associated with HCC proliferation (large tumor size). K-M survival analysis exhibited that the OS and relapse-free survival (RFS) of HCC patients with highly expressed miR-1275 were higher than those of those with a low expression level (p-value < 0.05) (Figure 1M,N). In order to explore the effects of miR-1275 on Hep-G2 and HCC-LM3 cells, we overexpressed and knocked down miR-1275 separately, and the efficiency of transfection in Hep-G2 cells was measured by qRT-PCR (Figure 1O,P). Moreover, the expression level of miR-1275 in Hep-G2 and HCC-LM3 cells co-treated with miR-1275 mimic and genistein was significantly higher than that in other groups (Figure 1Q,R). Thus, genistein can inhibit the viability of HCC cells and raise the expression level of miR-1275 in HCC cells.
EMT and stemness promote the progression and metastasis of HCC cells. In order to determine the effects of genistein and miR-1275 on the EMT and stemness of Hep-G2 and HCC-LM3 cells, we performed a series of in vitro functional experiments. In the proliferative experiment, the CCK-8 assay and EDU test were performed and indicated that the overexpression of miR-1275 or genistein treatment can restrain the proliferation of Hep-G2 and HCC-LM3 cells. Moreover, this inhibitory effect was further enhanced when combined with miR-1275 mimic and genistein (Figure 2A,B). Similarly, the same result was obtained from the motility experiments. The scratch-healing test and transwell migration assay were conducted and exhibited that the combination of miR-1275 mimic and genistein can further suppress the migration of Hep-G2 and HCC-LM3 cells (Figure 2C–I). Then, the spheroid formation test was applied to assess the stemness of Hep-G2 and HCC-LM3 cells, and it proved that miR-1275 mimic and genistein can synergistically attenuate the stemness of Hep-G2 and HCC-LM3 cells (Figure 2J–L). In order to better simulate the invasion of tumor cells, we performed a 3D spheroid invasion assay and discovered that miR-1275 mimic and genistein can also cooperate to inhibit the invasion of Hep-G2 and HCC-LM3 cells (Figure 2M–O). In addition, the Comet Assay was conducted to prove that the synergistic effect of miR-1275 mimic and genistein significantly aggravated DNA damage in Hep-G2 and HCC-LM3 cells and accelerated their apoptosis (Figure 3A,B). The flow cytometry analysis on Hep-G2 and HCC-LM3 cells also confirmed that miR-1275 mimic and genistein cooperatively decreased the proportion of S phase and increased the proportion of G0/G1 phase to further limit the proliferation of Hep-G2 and HCC-LM3 cells (Figure 3C,D). Finally, the expression levels of EMT and stemness markers in Hep-G2 and HCC-LM3 cells in four groups were measured by WB analysis. The results exhibited that, except for the increased expression of E-cadherin, the EMT markers (N-cadherin and Vimentin) and the stemness markers (SOX2, BMI1 and OCT4) in Hep-G2 and HCC-LM3 cells were declined by genistein and miR-1275 mimic (Figure 3E). Combined with previous qRT-PCR analysis results, it was clear that, under the synergistic effect of miR-1275 mimic and genistein, miR-1275 was significantly raised and further weakened the EMT and stemness of Hep-G2 and HCC-LM3 cells. Therefore, these in vitro functional experiments proved that miR-1275 upregulated by genistein can suppress HCC progression by inhibiting the EMT and stemness of HCC cells.
To explore the target gene of miR-1275, we obtained the intersection of four target gene prediction databases (TargetScan, miRWalk, miRDB and miRPathDB) of miR-1275 online, and the Venn diagram was also mapped (Figure 3F). However, due to the large number of predicted genes, we selected the genes that were highly expressed (threshold, p-value < 0.05 and log2FoldChange > 1) in HCC tissues in the TCGA-LIHC database and the Gene Expression Omnibus (GEO) database (GSE101685). There were 13 potential target genes of miR-1275 that qualified (Figure 3G). After further screening, we found that only EIF5A2 met the conditions. To confirm our conjecture, the expression levels of EIF5A2 in 70 pairs of HCC and matched adjacent tissues were analyzed by qRT-PCR, and it indicated that EIF5A2 was more highly expressed in HCC tissues than it was in peritumor tissues (Figure 3H). Meanwhile, Pearson’s correlation analysis showed that miR-1275 and EIF5A2 were negatively correlated in HCC tissues (n = 70). The Pearson’s correlation coefficient (r) and p-value (P) were also shown (Figure 3I). The protein levels of EIF5A2 were also found to be highly expressed in HCC tissues compared with adjacent tissues (n = 3) (Figure 3J). To determine the binding site of the 3′UTR region of EIF5A2 and miR-1275, the predicted binding site acquired from the TargetScan database and the mutant sequence of 3′UTR of EIF5A2 were listed (Figure 3K). A subsequent dual luciferase reporter test also confirmed that the luciferase activity of HEK293T cells co-transfected with miR-1275 mimic and EIF5A2-WT-3′UTR was lower than that in other groups (Figure 3L). Therefore, we proved that EIF5A2 was highly expressed in HCC tissues, and it was the target gene of miR-1275 in HCC cells.
To explore whether miR-1275 and genistein can exert the same effect in in vivo experiments, we constructed a xenograft tumor model and lung metastasis model of HCC-LM3 cells in nude mice. The transfection efficiency of ov-miR-1275 lentivirus in HCC-LM3 cells was shown, as follows (Figure S1C). The Xenograft HCC model showed that the volumes and weights of tumors in the genistein treatment group and the miR-1275 overexpression group (ov-miR-1275) were significantly smaller than those in the control group (ov-vector). Furthermore, the volumes and weights of tumors in the ov-miR-1275 + genistein group were the smallest (Figure 4A–C). The subsequent qRT-PCR analysis of miR-1275 expression levels in tumors proved that the miR-1275 in tumors of the ov-miR-1275 + genistein group was expressed more highly than that in other groups (Figure 4D). These results were consistent with previous in vitro assays. On the contrary, the immunohistochemical analysis of xenografts exhibited that the expression level of EIF5A2 in tumors of the ov-miR-1275 + genistein group was lower than that in other groups. In the same way, the expression levels of proteins (PI3K, p-Akt, Vimentin, SOX2 and Ki-67) in the ov-miR-1275 + genistein group were also the lowest (Figure 4E). In addition, genistein treatment and the overexpression of miR-1275 decreased the number of metastatic nodules in lung tissues, and the inhibition was the strongest in the ov-miR-1275 + genistein group (Figure 4F–H). These in vivo experiments proved that miR-1275, upregulated by genistein, can suppress the progression and metastasis of HCC by inhibiting EIF5A2.
EIF5A2 is a key gene known to regulate the EMT and stemness of various tumors [21,22]. In order to explore the common downstream signaling pathway regulated by miR-1275 and EIF5A2 in HCC cells, we performed bioinformatics analysis on EIF5A2 and miR-1275. First, based on the previous transcriptome sequencing results, we performed KEGG pathway enrichment analysis (threshold, p-value < 0.05) on the top 18 miRNAs with obvious differential expression levels. The signaling regulation network was constructed by DIANA-miRPath v3.0 database and R software. We found that miRNAs regulated by genistein mainly dominated the downstream PI3K-Akt signaling pathway (Figure S1A). Then, a pan-cancer analysis of EIF5A2 in the TCGA database indicated that EIF5A2 was highly expressed in various tumor tissues, including HCC (Figure 5A). According to the K-M survival analysis, the OS and RFS of HCC patients with highly expressed EIF5A2 in the TCGA-LIHC database were lower than those of those with lowly expressed EIF5A2 (Figure 5B,C). Meanwhile, the GSVA analysis of EIF5A2 in the TCGA database was conducted, and the heatmap was mapped by the EMTome database (Figure 5D). The GSVA analysis of EIF5A2 and the screened related signaling pathways (threshold, p-value < 0.05) in the TCGA-LIHC database were also shown in the chart. Interestingly, EIF5A2 was positively correlated with the PI3K-Akt signaling pathway (Figure 5E). Moreover, EIF5A2 was found to be positively correlated with EMT and the stemness markers of HCC in the TCGA-LIHC database (Figure S1D and Figure 5F). Therefore, we suspected that miR-1275 may attenuate the EMT and stemness of HCC cells by inhibiting the EIF5A2/PI3K/Akt signaling pathway. Finally, immunohistochemical staining and WB analyses confirmed that the expression levels of EIF5A2, PI3K and p-Akt declined evidently in Hep-G2 and HCC-LM3 cells treated with genistein miR-1275 mimic, but Akt changed inapparently (Figure S1E,F and Figure 4E). Thus, it can be concluded that miR-1275 upregulated by genistein can suppress the EMT and stemness of HCC cells by inhibiting the EIF5A2/PI3K/Akt signaling pathway.
In order to determine whether EIF5A2 can enhance the EMT and stemness of HCC, HCC-LM3 cells were co-transfected with miR-1275 mimic and EIF5A2 overexpression or EIF5A2-NC plasmid. At the same time, the HCC-LM3 cells in the four groups were all treated with genistein. The CCK-8 and EDU tests suggested that the proliferation ability of HCC-LM3 cells transfected with EIF5A2 overexpression plasmid was stronger than that in the EIF5A2-NC group and miR-1275 mimic group (Figure 5G–I). The scratch-healing and transwell migration assays also exhibited that the inhibitory effect of miR-1275 mimic on the migration of HCC-LM3 cells was partially weakened by EIF5A2 overexpression plasmid but not by EIF5A2-NC plasmid (Figure 5J–M). Moreover, in the spheroid formation and 3D spheroid invasion assays, the stemness and invasion of HCC-LM3 cells in the miR-1275 mimic+ EIF5A2 group were enhanced to some extent compared with that in the miR-1275 mimic + EIF5A2-NC group or miR-1275 mimic group (Figure 5L,N,O). The Comet Assay indicated that the DNA damage of HCC-LM3 cells in the miR-1275 mimic + EIF5A2 group was alleviated compared with that in the miR-1275 mimic + EIF5A2-NC or miR-1275 mimic groups (Figure 6A,B). The flow cytometry analysis found that, compared with the miR-1275 mimic + EIF5A2-NC and miR-1275 mimic groups, the proportion of the S phase was raised and the proportion of the G0/G1 phase was reduced in the miR-1275 mimic+ EIF5A2 group (Figure 6C,D). Finally, the WB analysis found that the expression level of EIF5A2 in the miR-1275 mimic+ EIF5A2 group was improved compared with that in the miR-1275 mimic or miR-1275 mimic + EIF5A2-NC group, but it was still lower than that in the NC group. In addition, compared with other groups, the expression levels of PI3K, p-Akt, EMT markers (N-cadherin and Vimentin) and stemness markers (SOX2, BMI1 and OCT4) in the miR-1275 mimic+ EIF5A2 group had the same trend as those in the EIF5A2 group. However, the expression level of the EMT marker (E-cadherin) in the miR-1275 mimic+ EIF5A2 group was the opposite, and there was no significant difference in the expression levels of Akt among the four groups (Figure 6E). From this, we drew a conclusion that the inhibitory effect of miR-1275 on the EMT and stemness of HCC can be reversed by the activation of the EIF5A2/PI3K/Akt signaling pathway. The pattern diagram of interpreting the miR-1275/EIF5A2/PI3K/Akt axis in HCC-LM3 cells was exhibited (Figure 6F). Therefore, it was clear that genistein can positively regulate miR-1275 to suppress the EMT and stemness of HCC cells by inhibiting the EIF5A2/PI3K/Akt signaling pathway in vivo and vitro.
Genistein (C15H10O5, Mr = 270.24) accounts for 60% of soybean isoflavones (SIF) in soybeans [23]. In previous studies, genistein was mainly focused on for its role in inhibiting tumor growth or inducing apoptosis. Similarly, it has been well demonstrated that genistein has the ability to inhibit HCC progression and induce HCC apoptosis [13,24]. At the initial stage of our study, we determined that the inhibitory effects on HCC viability were dependent both on the time and the doses of genistein treatment, and the results were consistent with other previous studies [8,9,13]. On these grounds, we raised the question of through which way genistein can regulate HCC progression and metastasis. Then, we paid attention to the EMT and stemness of HCC. On the one hand, the occurrence of EMT leads to a decrease in cancer cells adhesion and an enhancement of their invasion and metastasis. On the other hand, the enhancement of stemness makes cancer cells possess powerful self-renewal and multidirectional differentiation abilities [3,25]. Briefly speaking, EMT and stemness are indispensable in promoting the progression and metastasis of cancers. Next, we further speculated whether genistein can exert an anti-tumor effect by abating the stemness and EMT of HCC cells. In addition, it has also been widely reported that miRNAs play an important role in dominating the EMT and stemness of HCC [2,26]. For example, miR-192-5p and miR-568 were found to regulate the stemness behavior to affect HCC progression [27,28]. In our study, we found for the first time that miR-1275 can be positively regulated by genistein in HCC cells through transcriptome sequencing and series experiments. Before our study, the miRNA array analysis of HCC tissues performed by Wen Wang et al. [29] and other studies also supported that miR-1275 is involved in inhibiting the progression and metastasis of HCC [30]. As indicated earlier, miR-1275 is essential for genistein to inhibit HCC progression and metastasis. In addition, the EMT and stemness of HCC cells were indeed weakened by miR-1275 upregulation according to the results of immunohistochemistry and WB. It proved that miR-1275 abated the EMT and stemness to further suppress the progression and metastasis of HCC. It is a priority to figure out the key gene and common signaling pathway between EMT and stemness in controlling HCC progression and metastasis. Through prediction by databases and the dual luciferase reporter test, EIF5A2 was found to have potential to control the stemness and EMT of HCC. EIF5A2 (eukaryotic initiation factor 5A2) is located at 3q26 on human chromosomes [21]. EIF5A2 can act as a transcription factor to drive the occurrence of multiple cancers and induce poor prognosis [21]. EIF5A2 has the effect of enhancing the EMT and stemness of tumors, which promotes the progression, recurrence and metastasis of cancers [31,32]. For example, it was reported that EIF5A2 was involved in maintaining the existence of cancer stem cells in HCC cells through the c-Myc pathway [32]. In addition, Zhiyuan Zhang et al. [31] confirmed that EIF5A2 can improve the EMT to promote the metastasis of colorectal cancer. Overall, EIF5A2 is the key gene that simultaneously regulates the interaction between EMT and stemness. Meanwhile, there are many common signaling pathways between EMT and stemness, such as PI3K/Akt, Wnt/β-catenin and TGF-β [33,34,35]. Because of this, through KEGG and GSVA enrichment analyses, we realized that EIF5A2/PI3k/Akt was the key signaling pathway regulating HCC EMT and stemness. PI3K (phosphatidylinositol kinase) is a dimer composed of regulatory subunit p85 and catalytic subunit p110. The upregulation of PI3K can change the structure of Akt and activate it by phosphorylation, thus regulating cell proliferation, differentiation, apoptosis, migration and other phenotypes [34,36]. The PI3K/Akt pathway has proven that it can be activated by the degradation of P85a or RALYL to enhance the stemness and promote the EMT of HCC [3,36]. The knockdown of EIF5A2 can restrain the PI3K/Akt signaling pathway to prevent the growth of bladder cancer [37]. Subsequent WB and immunohistochemical analysis also exhibited that EIF5A2/PI3K/Akt attenuated by miR-1275 and genistein suppressed the EMT and stemness of HCC. This inhibitory effect induced the retardation of HCC progression and metastasis. Moreover, the DNA damage of HCC cells was aggravated, and cell cycles were also blocked in the GO/G1 phase. In view of this, miR-1275 upregulated by genistein can indeed inhibit the EIF5A2/PI3K/Akt pathway to suppress the EMT and stemness of HCC. Despite all this, there were still some deficiencies in this study. First, the concentrations of genistein used in our study were consistent with the recommended concentration range (2.7–27.0 µg/mL) [8,9,13]. However, several scholars pointed out that different doses of genistein have different effects on tumors, and the specific drug concentration of genistein is still controversial [38,39,40]. In the follow-up study, we will further explore the anti-tumor mechanisms of genistein at different concentrations. Second, the tumor microenvironment (TME) is also very important to the crosstalk between EMT and stemness, and we will focus on the effects of genistein on the communication between HCC cells and other cells in TME. In summary, genistein has powerful anti-HCC effects and can play a variety of roles to inhibit the progression and metastasis of HCC. In our study, we confirmed that miR-1275 upregulated by genistein can attenuate HCC progression and metastasis by suppressing the EIF5A2/PI3K/Akt signaling pathway to inhibit the EMT and stemness of HCC cells. It is clear that genistein can control the expression levels of miRNAs to restrain HCC. However, the dose and duration of genistein are still controversial, and it will take more time to be verified. Our research provides a new direction and theoretical basis for seeking new adjuvant drugs for HCC, and miR-1275 may become a new therapeutic target of HCC.
In summary, our study indicated that miR-1275 upregulated by genistein can attenuate the EMT and stemness of HCC by inhibiting the EIF5A2/PI3K/Akt signaling pathway and thus restrict HCC progression and metastasis. | true | true | true |
|
PMC9598940 | Caterina Russo,Alessandro Maugeri,Laura De Luca,Rosaria Gitto,Giovanni Enrico Lombardo,Laura Musumeci,Giovambattista De Sarro,Santa Cirmi,Michele Navarra | The SIRT2 Pathway Is Involved in the Antiproliferative Effect of Flavanones in Human Leukemia Monocytic THP-1 Cells | 24-09-2022 | flavonoids,SIRT2,naringenin,hesperetin,naringin,neohesperidin,leukemia,cancer,flavanones,citrus | Acute myeloid leukemia (AML) represents the most alarming hematological disease for adults. Several genetic modifications are known to be pivotal in AML; however, SIRT2 over-expression has attracted the scientific community’s attention as an unfavorable prognostic marker. The plant kingdom is a treasure trove of bioactive principles, with flavonoids standing out among the others. On this line, the aim of this study was to investigate the anti-leukemic properties of the main flavanones of Citrus spp., exploring the potential implication of SIRT2. Naringenin (NAR), hesperetin (HSP), naringin (NRG), and neohesperidin (NHP) inhibited SIRT2 activity in the isolated recombinant enzyme, and more, the combination between NAR and HSP. In monocytic leukemic THP-1 cells, only NAR and HSP induced antiproliferative effects, altering the cell cycle. These effects may be ascribed to SIRT2 inhibition since these flavonoids reduced its gene expression and hampered the deacetylation of p53, known sirtuin substrate, and contextually modulated the expression of the downstream cell cycle regulators p21 and cyclin E1. Additionally, these two flavanones proved to interact with the SIRT2 inhibitory site, as shown by docking simulations. Our results suggest that both NAR and HSP may act as anti-leukemic agents, alone and in combination, via targeting the SIRT2/p53/p21/cyclin E1 pathway, thus encouraging deeper investigations. | The SIRT2 Pathway Is Involved in the Antiproliferative Effect of Flavanones in Human Leukemia Monocytic THP-1 Cells
Acute myeloid leukemia (AML) represents the most alarming hematological disease for adults. Several genetic modifications are known to be pivotal in AML; however, SIRT2 over-expression has attracted the scientific community’s attention as an unfavorable prognostic marker. The plant kingdom is a treasure trove of bioactive principles, with flavonoids standing out among the others. On this line, the aim of this study was to investigate the anti-leukemic properties of the main flavanones of Citrus spp., exploring the potential implication of SIRT2. Naringenin (NAR), hesperetin (HSP), naringin (NRG), and neohesperidin (NHP) inhibited SIRT2 activity in the isolated recombinant enzyme, and more, the combination between NAR and HSP. In monocytic leukemic THP-1 cells, only NAR and HSP induced antiproliferative effects, altering the cell cycle. These effects may be ascribed to SIRT2 inhibition since these flavonoids reduced its gene expression and hampered the deacetylation of p53, known sirtuin substrate, and contextually modulated the expression of the downstream cell cycle regulators p21 and cyclin E1. Additionally, these two flavanones proved to interact with the SIRT2 inhibitory site, as shown by docking simulations. Our results suggest that both NAR and HSP may act as anti-leukemic agents, alone and in combination, via targeting the SIRT2/p53/p21/cyclin E1 pathway, thus encouraging deeper investigations.
Acute myeloid leukemia (AML) is the most common hematologic neoplasm in adults, characterized by an accumulation of abnormal myeloblasts, most frequently in the bone marrow, leading to its failure and, eventually, death [1]. The incidence of AML is greater for male subjects than female ones, and it was reported to be increased in more developed and economically advanced countries [2]. Despite the acknowledged role of genetic aberrations in the devolvement of AML, its exact etiology still remains elusive [3]. Therefore, scientific community constantly seeks for novel molecular targets to design more effective therapies in counteracting this nefarious disease. The acetylation state of histones regulates epigenetic processes by affecting transcription factor access to DNA, thus determining gene expression levels. Histone deacetylases (HDACs) are implied in this mechanism, which in turn increases DNA/histone complex compaction [4,5]. Among HDACs, human sirtuins, NAD+-dependent enzymes, are the focus of scientific community’s attention owing to the fact that they regulate a large number of cellular pathways implied in cellular aging and age-associated diseases, among which cancer [6]. In particular, SIRT2, one of the seven human sirtuins, is mainly a cytoplasmic enzyme involved in the deacetylation of histones and α-tubulin, as well as many other transcriptional factors (i.e., p53 and NF-κB). Interestingly, elevated mRNA levels of SIRT2 are found in AML patients’ blasts compared to those of healthy subjects. Specifically, it is over-expressed in both intermediate- and poor-risk patients, compared to the favorable-risk ones, as well as being associated with a significantly shorter overall and event-free survival rate. Moreover, SIRT2 over-expression was particularly evident in AML patients belonging to the M5 subtype, according to the French-American-British (FAB) classification of AML, which is defined as acute monocytic leukemia [7]. Therefore, compounds able to inhibit SIRT2 activity are thought to be novel therapeutical approaches to ameliorate conditions such as leukemias, among which AML. Plant kingdom provides an uncountable number of active principles capable of modulating, blocking, or enhancing several cellular pathways, involved in a wide plethora of physio-pathological conditions. Flavonoids stand out among the others for their anti-infective [8], neuroprotective [9,10], antioxidant and anti-inflammatory [11,12,13] and anti-cancer activities [14,15,16,17]. Moreover, it was shown that flavonoids, such as quercetin and derivatives, have already proved their potential in inhibiting SIRT2 [18]. On this line, the aim of this study was to investigate whether the main flavanones of Citrus spp., namely naringenin (NAR), hesperetin (HSP), naringin (NRG) and neohesperidin (NHP; Figure 1), are able to inhibit the activity of SIRT2 enzyme in both cell-free and in vitro models, thus investigating their anti-leukemic activity.
SIRT2 activity was determined using SIRT2 Direct Fluorescent Screening assay kit (n. 700280) purchased from Cayman Chemical Company (Ann Arbor, MI, USA) [19]. For cell-free studies, the histone deacetylase activity was assayed using SIRT2 human recombinant enzyme provided by the kit. Increasing concentrations of HSP, NAR, NRG, and NHP (10, 50, 100, 200, 400 µM; Sigma-Aldrich, Milan, Italy) were tested according to the manufacturer’s guidelines. SIRT2 inhibitor SirReal2 (140 nM; Selleckchem, Houston, TX, USA) was used as positive control. The fluorescence was read after 30 min using a FLUOstar Omega Plate Reader (BMG LABtech, Ortenberg, Germany) at 350–360 nm excitation wavelength and 450–465 nm emission wavelength.
The effects of the combination between increasing concentrations of HSP and NAR (50–400 µM) was assessed in terms of SIRT2 activity on the recombinant isolated enzyme, combined at different ratios, following the checkboard method. In particular, pharmacological interaction models of Loewe additivity and Bliss independence have been developed, and a recent protocol has been employed to measure the Zero Interaction Potency (ZIP) [20]. Results were processed to define interaction between the abovementioned compounds. Synergy scoring was determined using the SynergyFinder 2.0 software [21] that exploits the ZIP calculation method, expressing the synergism as δ score. Positive δ values correspond to synergism, whereas negative ones to antagonism [22].
The human leukemia monocytic THP-1 cell line was originally obtained from ATCC (Rockville, MD, USA). The cells were grown in RPMI 1640 medium supplemented with 10% v/v heat-inactivated fetal bovine serum (FBS), L-glutamine (2 mM), HEPES (10 mM), sodium pyruvate (1 mM), glucose (2.5 g/L), 2-mercaptoethanol (0.05 mM), penicillin (100 IU/mL) and streptomycin (100 µg/mL), at 37 °C in a humified 5% CO2 atmosphere. Medium was renewed every 2 days and split performed when cells reached maximum density (1 × 106 cells/mL). Human peripheral blood mononuclear cells (PBMCs) were originally obtained from ATCC and cultured using supplemented RPMI 1640 medium (10% v/v FBS), at 37 °C in a 5% CO2 environment. Each reagent for cell culture was from Gibco (Life Technologies, Monza, Italy).
The evaluation of the antiproliferative activity was assessed by resazurin assay, as an index of mitochondrial functionality, and propidium iodide (PI) staining to detect dead cells [23]. For the former, THP-1 cells and PBMCs were seeded in 96-well plates at a density of 1 × 104 cells/well in 200 µL. Cells were treated with the flavanones NAR, NRG, HSP, NHP (50–400 µM). For vehicle control, we added the same amount of DMSO present in the highest concentration of the flavanones tested to assess whether no toxic effect was induced by the solvent. The plates were then incubated for 24, 48 and 72 h at 37 °C, prior adding 20 µL of resazurin (Santa Cruz, Dallas, TX, USA) 0.01% w/v solution to each well and kept for further 3 h. Fluorescence was measured on a microplate reader POLARstar Omega (BMG Labtech) with an excitation wavelength of 544 nm and an emission one of 590 nm. For PI staining, cells (1 × 106 cells/well) were treated with NAR, NRG, HSP, NHP (100–400 µM) in 6-well plates, which were incubated at 37 °C for 72 h. Subsequently, cells were harvested, washed twice and resuspended in 100 µL of PBS plus 10 µL of PI solution (10 µg/mL; Sigma-Aldrich). Stained cells were incubated for 30 min at room temperature in the dark and fluorescence of at least 10,000 events was analyzed by a Novocyte 2000 cytofluorimeter (Agilent, Santa Clara, CA, USA) with FL-2 channel.
The ability of flavanones to interfere with progression of cell cycle was evaluated by flow cytometry [24]. Briefly, THP-1 cells were seeded in 6-well plates (2 × 105 cells/well) and treated with the flavanones NAR and HSP (100–400 µM) or with their combination at concentrations of 100 µM for 24, 48 and 72 h. Then, cells were collected, centrifuged, washed with PBS, and fixed in 70% ice-cold ethanol while gently vortexed. After at least 2 h at 4 °C, cells were centrifugated, washed twice with cold PBS, and resuspended in 250 μL of PBS together with 5 μL of RNase A (10 mg/mL; Sigma-Aldrich) at 37 °C for 1 h. After incubation, 10 µL of PI (1 mg/mL) were added to samples, and immediately acquired by cytofluorimeter. Three independent sets of at least 10,000 events were collected for each condition.
To assess the levels of acetylated p53 in THP-1 cells treated for 24 h with NAR and HSP (100, 200 and 400 µM) and their combination (100/100 µM), or with SIRT2 inhibitors SirReal2 (10 µM) and nicotinamide (NAM; 1 mM; Cayman, Ann Arbor, MI, USA) [25], a commercial enzyme-linked immunosorbent assay (ELISA) kit was employed (E4531; Biovision, Milpitas, CA, USA). Briefly, cell lysates from the abovementioned treatments were quantified using Bio-Rad DC Protein Assay (Bio-Rad Laboratory, Hercules, CA, USA) with bovine serum albumin as standard. To equal amounts of protein for each sample, 100 µL of biotin-conjugated primary antibody were added in provided strips and incubated for 1 h at 37 °C. After washing wells, 100 µL of streptavidin HRP-conjugated were added and incubated for additional 30 min at 37 °C. Finally, 90 μL of TMB substrate were added into each well, cover the plate and plate was incubate at 37 °C in dark for further 30 min. Color formation was stopped and absorbance recorded with a microplate spectrophotometer at 450 nm (iMark™ microplate reader, Bio-Rad Laboratories). Results were extrapolated as ratio between values detected in treated and untreated cells.
THP-1 cells were plated in 100 mm Petri dishes at a density of 1 × 106 cells/dish with fresh medium (untreated cells), NAR and HSP (100, 200 and 400 µM), as well as their combination (100/100 µM) at 37 °C for 12 and 24 h. Then, total RNA from untreated and treated cells was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions. Afterward, 2 µg of total RNA was reverse transcribed into cDNA using the High-Capacity cDNA Archive Kit (Applied Biosystems, Life Technologies, Foster City, CA, USA), as previously described [26]. Quantitative PCR (qPCR) reactions were performed in triplicate into a 96-well plate, in a final volume of 20 µL, containing 1× SYBR Select Master Mix (Applied Biosystems), 0.2 µM of specific primers and 25 ng of RNA, previously converted into cDNA. The mRNA levels of SIRT2, p53 (TP53), p21 (CDKN1) and cyclin E1 (CCNE1) were analyzed using a 7500 qPCR System (Applied Biosystems), according to the following protocol: one cycle at 95 °C for 10 min, followed by 45 cycles at 95 °C for 15 s, and 60 °C for 1 min. A standard dissociation stage was added to assess the primer specificity. The primer sequences used for qPCR were designed based on those previously published and are listed in Table 1. Data collected were analyzed employing the 2−∆∆CT relative quantification method against β-actin (ACTB), used as the housekeeping control. The values are expressed as n-fold change compared to untreated cells. When the value was less than 1, it was converted into its inverse to report downregulated genes.
Docking studies were performed by AutoDock4 suite [27], using the crystal structure of SIRT2 in complex with SirReal2 retrieved from the RCSB Protein Data Bank (PDB code 4RMG) [25]. The ligand and water molecules were removed, and hydrogens were added by Discovery Studio 2.5. Ligand structures were generated by VEGAZZ suite and optimized by following a conjugate gradient minimization by AMMP calculation, implemented in the VEGAZZ program [28]. Docking simulation was performed by following the same protocol as reported by Roshdy et al. [29], except for the grid center’s coordinates that were defined using the centroid of the co-crystalized ligand SirReal2.
The assays were carried out in different replicates, as described above. Data obtained were expressed as mean ± standard error of the means (SEM). Statistical evaluation of results was performed using one-way analysis of variance (ANOVA), depending on the assay. Multiple comparisons of the means of the groups were performed by the post hoc Student–Newman–Keuls test (SigmaPlot Software, Chicago, IL, USA). The p-values lower or equal to 0.05 were considered statistically significant.
The inhibition of SIRT2 enzymatic activity induced by the flavanones NAR and HSP, along with their glycosidic counterparts NRG and NHP, respectively, was assayed in a cell-free model consisting of the isolated recombinant enzyme. In this setting, each flavanone was able to inhibit SIRT2 activity, despite to a different extent (Figure 2). In detail, at 200 µM, the two aglycones HSP and NAR reduced the enzymatic activity up to 52.3 ± 4.1% and 44.8 ± 3.2%, respectively, whereas, at 400 µM, up to 65± 4.7% and 63.8 ± 3.5%, respectively (Figure 2A,B). On the contrary, NHP and NRG were the weakest among the flavanones tested, reaching an inhibition of 38.2 ± 2.1% and 39.2 ± 4.4% at 200 µM, and up to 47.1 ± 3.2% and 51 ± 4.8% at 400 µM, respectively (Figure 2C,D).
SIRT2 enzymatic activity was further evaluated by testing the combination of single flavanones (NAR, NRG, HSP, and NHP) at different ratios in the isolated enzyme. The obtained results have been processed to investigate the presence of antagonism or synergism. A clear antagonism was observed between the pairs NAR-NHP, NAR-NRG, NHP-NRG, and NRG-HSP, whereas a weak synergistic interaction was detected between NHP and HSP (data not shown). Unlike the other combinations, the flavanones NAR and HSP, which individually elicited the strongest inhibitory effect on SIRT2 enzymatic activity, showed a great synergism in counteracting SIRT2 activity (Figure 3). Indeed, the combination between NAR and HSP displayed an overall sharp synergistic effect with a ZIP score (δ) of 8.076. As showed, the red area (synergism) reaches a peak when the two flavanones are at equimolar concentrations of 100 µM (δ = 12.119). Moreover, the synergism of the two flavanones appears being preserved for equimolar concentrations up to 200 µM (δ = 8.819) and weakens toward higher concentrations. Therefore, the effective combinations of the two flavanones at equimolar concentrations of both 100 and 200 µM were then assessed in vitro.
To investigate whether the studied flavanones possessed anti-leukemic activity, the effect of NAR and HSP, along with their glycosidic counterparts NRG and NHP, was assessed in a human monocytic leukemia THP-1 cell line, which represents a preclinical model of the M5 subtype of AML. In these cells, both NRG and NHP hampered viability at 72 h by no more than 20% at the highest concentration tested (400 µM). On the contrary, the aglycones NAR and HSP significantly reduced cell proliferation already at 24 h (29.0 ± 2.4% and 35.0 ± 2.2%, respectively) and up to 64.0 ± 2.3% and 66.0 ± 2.4% at 72 h, respectively (Figure 4A). In order to assess the potential cytotoxicity of the flavanones under study, we repeated the resazurin assay in normal human PBMCs, testing the same timings and concentrations. Noteworthy, none of these caused any inhibitory effects on PBMCs growth, if not after exposure to the flavanones NAR and HSP at the concentration of 400 µM for 72 h, where a slight but not significant reduction in cell viability was observed (Figure 4B). At this point, we have thus identified non-cytotoxic concentrations employed in the following experiments. The PI staining corroborated the outcome observed with the resazurin assay (Figure 5). Consistent with the latter, both NAR and HSP induced cell death in THP-1 monocytes, as demonstrated by the increase in the number of fluorescent cells by 64.5 ± 1.9% and 66.1 ± 3.1% at 72 h, respectively. As for cell viability assay, neither NRG nor NHP was able to induce any damage to the THP-1 cell membrane; thus, no PI diffused up to the nuclear compartment, thus intercalating to DNA (Figure 5A). On the contrary, no flavanone caused significant cytotoxicity against primary cells after 72 h of exposure (Figure 5B). Given these results, further in vitro experiments were performed considering only NAR and HSP.
With the aim of comprehending the mode of death elicited by NAR, HSP, and their combination, their influence on the cell cycle progression of THP-1 cells was evaluated. After 24 h of treatment, just a slight modulation of cell cycle progression was observed with 400 µM of NAR and HSP (data not shown). After 48 h of treatment, 100 µM of NAR did not elicit any modification of the ratio among cell cycle phases respect to the control, whereas 200 µM and even more 400 µM increased the percentage of cells in S phase. Similarly, 100 µM of HSP did not alter the cell cycle at 48 h, while 200 µM increased the cell population in S and G2/M phases. The accumulation of cells in the S phase was also recorded with HSP at 400 µM concentration. Moreover, the combination between NAR and HSP at a molar ratio of 1:1 (100 µM) displayed a stronger effect than the two flavanones alone (Figure 6A). Consistent with the synergistic interaction observed in the abiotic assay (red area of Figure 3), the combination between NAR and HSP at a concentration of 200 µM produced an effect comparable to that from the 100 µM combination (data not shown). After 72 h of treatment, both NAR and HSP at 100 and 200 µM brought a sharp increase in cell population in the S phase, with HSP also in G2/M one. For both flavanones, the S phase arrest is even more appreciable at a concentration of 400 µM (Figure 6B). Remarkably, as for the 48-h treatment, the 100 µM combination between the two flavanones elicited a stronger effect than that recorded for single compounds. A similar block of the cell cycle as that of the 100 µM combination was also experienced with the 200 µM one (data not shown). Of note, cells populating the sub-G0/G1 phase (hypodiploid cells), a known sign of apoptosis, are relevant after treatment with HSP (all tested concentrations), NAR (400 µM), and their combination (100/100 µM) at both 48 and 72 h (Figure 6).
As a direct in vitro effect of SIRT2 activity modulation, the levels of acetylated p53 protein were evaluated by means of an ELISA. In detail, treatment with both NAR and HSP 100 µM for 24 h did not alter levels of acetylated p53 with respect to control cells (Figure 7). On the contrary, NAR 200 µM increased acetylation of p53 by 1.43-fold, while HSP at the same concentration by 1.27-fold, with respect to controls. In turn, NAR 400 µM increased p53 acetylation up to 1.64-fold as well as HSP 400 µM to 1.53-fold, compared to controls. Notably, the combination between NAR and HSP (100 µM) elicited an enhancement of the rate of p53 acetylation by 1.37-fold with respect to controls (Figure 7). Similarly, this effect occurred with a combination of 200 µM (data not shown). Finally, to appropriately compare the abovementioned results, we used a specific (SirReal2, 10 µM) and a non-specific SIRT2 inhibitor (NAM, 1 mM) as positive controls. In comparison with what occurs in untreated cells, the incubation of THP-1 cells with both inhibitors significantly increased the degree of p53 acetylation by 1.23- and 1.67-fold, respectively. This suggests that NAR and HSP behave as synthetic SIRT2 inhibitors.
Besides inhibiting SIRT2 enzymatic activity, NAR and HSP were also able to modulate its gene expression. Indeed, the treatment with both HSP 200 µM and NAR 400 µM for 12 h significantly reduced SIRT2 mRNA levels by 1.25-fold down compared to controls, as well as their equimolar combination of 100 µM concentration, which brought no effects for single flavanones. Of note, the highest tested concentration of HSP (400 µM) significantly lowered SIRT2 mRNA expression after 12 h (1.85-fold down vs. CTRL; Figure 8A). At 24 h of treatment, none of the flavanones was able to alter SIRT2 expression, except for HSP 400 µM, which decreased it by 1.32-fold down vs. CTRL (Figure 8A). Neither NAR nor HSP altered the gene expression of total p53 for concentrations up to 200 µM at any of the timings tested, maintaining the mRNA quantities at the level detected in untreated cells. However, a non-significant increase in p53 gene expression was observed at 400 µM of NAR and HSP after both 12 h and 24 h (Figure 8B). Moreover, the treatment of THP-1 cells with NAR and HSP was able to modulate S phase cell cycle-related factors p21 and cyclin E1, already after 12 h of treatment. In particular, neither NAR nor HSP 100 µM modified p21 expression with respect to control cells, whereas this occurred with higher concentrations of both flavanones (200 and 400 µM). In detail, compared to controls, treatment with NAR 200 µM for 12 h increased p21 mRNA levels in THP-1 cells by 1.7-fold and with HSP 200 µM by 1.5-fold. For the concentrations of 400 µM of NAR and HSP, a similar increase in p21 levels (1.9- and 1.7-fold, respectively) was recorded after 12 h. Likewise, this growing trend was maintained after 24 h of treatment with NAR (2.7-fold vs. CTRL) and HSP (2.2-fold vs. CTRL) 200 µM, whereas it was even higher with NAR 400 µM, which significantly raised p21 gene expression by 4.0-fold, and HSP 400 µM by 4.9-fold respect to controls. On this line, after 24 h of incubation, the equimolar combination (100 µM) between the two flavanones increased p21 mRNA levels by 2.1-fold respect control, similarly to HSP 200 µM (Figure 8C). Contrariwise, cyclin E1 expression was significantly hampered by NAR and HSP already after 12 h of treatment. Indeed, NAR reduced mRNA levels of cyclin E1 by 1.25-, 1.62- and 2.49-fold down at 100, 200, and 400 µM, respectively, whereas HSP lowered cyclin E1 mRNA quantity by 1.70- and 2.04-fold down at 200 and 400 µM, respectively. Their combination decreased by 1.52-fold down cyclin E1 expression after 12 h of incubation. After 24 h of treatment, NAR and HSP significantly reduced cyclin E1 gene expression at 100 µM (2- and 1.43-fold down, respectively), 200 µM (5- and 2.9-fold down, respectively), and 400 µM (6.7- and 5-fold down, respectively). Their association (100/100 µM) brought a result like that of the 200 µM of both NAR and HSP alone (Figure 8D), while the 200 µM combination did not prove any better than that reported (data not shown).
To investigate the molecular interactions between the flavanones NAR and HSP with SIRT2, we performed docking simulations using AutoDock4 suite [27], employing crystallographic coordinates of the complexed 4RMG as reference structure [25]. The docking protocol (see Methods) was first validated through self-docking of SIRT2 inhibitor SirReal2, whose best docking pose was in suitable agreement with the experimental structure (RMSD value of 0.5 Å). The computational studies suggested that all studied compounds had a similar network of interactions as SirReal2 in the lipophilic pocket of SIRT2, forming crucial contacts within active site residues. The inspection of docking results suggested that the 2-(4-hydroxyphenyl)-substituent of NAR, taken as an example, forms crucial π-T-shaped interaction with Phe131, whereas the chromen-4-one moiety occupies the hydrophobic region close to Tyr139 and Phe190 residues. Moreover, our molecular simulation revealed that NAR, such as HSP, made two hydrogen bond interactions with oxygen atoms of Ile118 and Ala135 backbone. Compared to SirReal2, NAR loses the crucial π-π stacking interaction within the selectivity pocket of SIRT2. This could explain the minor activity of the flavanones (Figure 9).
The main source of dietary flavonoids is Citrus fruits, and, due to their presence, these are endowed with well-acknowledged protective activities exploited to defend human health [30,31,32], also in clinical settings [33,34]. Regarding cancer, Citrus fruits and their derivatives have been long investigated for their role in this nefarious disease [35,36,37], and flavanones stand out among others [38]. In this work, we started from the assumption that quercetin-derived compounds have been demonstrated to possess SIRT2 inhibition activity [18], and, given the implication of this enzyme in AML [7], we investigated the role of SIRT2 in the anti-leukemic effect of the most common flavanones of Citrus spp. Therefore, the first step was to assess the effect of NAR, HSP, and their glycosidic counterparts, NRG and NHP, on the recombinant isolated SIRT2 enzyme. Interestingly, all flavanones were able to inhibit the deacetylase activity of SIRT2, despite to different extent, thus being the first to report these effects. Previously, we demonstrated how natural products combined may elicit stronger effects than when alone [22]. On this line, we assessed the pair combinations of the four flavanones employing the same abiotic assay as for testing the single compounds. Unlike all the other combinations, mainly exhibiting an antagonistic interaction, NAR and HSP synergistically inhibited SIRT2 activity, reaching a maximum at an equimolar ratio of 100 µM. The synergism between these two flavanones was maintained for equimolar concentrations up to 200 µM and then progressively decreased at higher concentrations. In the light of the inhibitory effects of the selected flavanones on the SIRT2 recombinant enzyme and that SIRT2 expression was increased in the M5 subtype of AML patients [7], we decided to assess their activity on a well-established cellular model of AML belonging to this subtype, as the human monocytic leukemic THP-1 cells [39]. This is a reliable cell line for functional, preclinical therapeutics, and target identification studies, as well as for investigating monocyte function and differentiation. Previous reports suggested the capability of many flavonoids as anti-leukemic agents via targeting multiple pathways [40,41]. In our study, we confirmed this assumption, enriching it with further details. Indeed, we demonstrated that NAR and HSP exerted interesting antiproliferative activity in THP-1 cells, whereas the glycosides NRG and NHP showed minimal effect on the proliferation of this cell line, at least at the concentration tested in our study. This outcome is in line with a previous work by Chen and co-workers who claimed that rutinoside in C-7 of flavonoids, the carbohydrate moiety characterizing both NRG and NHP, prevents the induction of antiproliferative activity [42]. Interestingly, these compounds did not exert any toxicity on PBMCs, suggesting a safety profile of flavanones employed in this study. Given these results, we investigated which type of cell death was elicited by NAR and HSP by assessing whether these compounds can interfere with the progression of THP-1 cells during the cell cycle. This cellular process is governed by checkpoints that are carefully followed by normal cells; nevertheless, if genetic aberrations or abnormalities occur, cells will enter a state of cell cycle arrest [43]. In our experiments, both NAR and HSP were able to block the cell cycle, increasing cells in the S phase at both 48 and 72 h. Notably, the association of the two flavanones retraced what we witnessed in the abiotic assay, where they synergistically acted to inhibit SIRT2 enzymatic activity. Moreover, the hypodiploid population increased in THP-1 cells treated with NAR and HSP 200 and 400 µM, as well as their association (100 µM each), indicating that the elicited block of the cell cycle can lead to apoptotic cell death. The pro-apoptotic activity of 200 µM of NAR in THP-1 cells was also reported by Park and co-workers [44] that linked this property to the activation of caspases and mitochondrial dysfunctions, associated with the inactivation of the PI3K/AKT pathway. Interestingly, we are the first to report the induction of apoptosis elicited by HSP in this cell line. In addition, our results are in line with those of Chen and co-workers [42], who reported that low concentrations of NAR and HSP (i.e., 20, 40, and 80 µM), as well as of their corresponding aglycones, did not exert apoptosis in THP-1 cells, contrariwise to HL-60 ones. To verify the sirtuin inhibition by flavanones in vitro, we evaluated the level of p53 acetylation, the main cellular target of SIRT2 deacetylase activity [45]. Interestingly, both NAR and HSP 200 and 400 µM hampered p53 deacetylation in a significant manner, like the equimolar association of the two at a lower concentration (100 µM), further corroborating our initial hypothesis of synergism. In this setting, SirReal2, a specific SIRT2 inhibitor, altered p53 acetylation such as NAR and HSP, thus supporting their capability of inhibiting SIRT2 deacetylase activity in vitro. On the other hand, treatment of THP-1 cells with NAM, a non-specific SIRT2 inhibitor, determined a considerable increase in acetylated p53 levels. This is due to a combined effect on more than one sirtuin, which, similarly to the NAM, our flavanones seem to exert [19]. Proving this, the selectivity of NAR and HSP (200 and 400 µM) toward the SIRT2 enzyme seems to straddle that of SirReal2 (10 µM) and NAM (1 mM). However, we saw no change in p53 gene expression for none of the treatments. Since the enzyme modulation can be achieved by either affecting the activity or the gene expression, we wondered whether the lowered SIRT2 activity could be originated by a decrease in its mRNA levels, in addition to the inhibition of its enzymatic activity. Therefore, when we quantified the total SIRT2 gene levels, we observed an early modulation (12 h) of mRNA expression at the highest tested concentration (400 µM) of NAR and HSP, as well as with their combination at equimolar ratio (100 µM), an effect that was totally abolished at 24 h of treatment, except for HSP 400 µM. This outcome may suggest that cancer cells respond to SIRT2 inhibition with an increase in its gene expression in order to restore its functionality, which is crucial for the survival of AML cells. Notably, HSP 400 µM steadily maintained the suppression of SIRT2 expression, suggesting its potentiality in AML. During the cell cycle, p53 directly binds DNA to promote the transcription of several factors, among which p21 is the pivotal one [46]. The immediate effect of the p21 increase is the regulation of cyclins such as D and E [47]. In particular, p21 activation leads to inhibition of the cyclin E1, which in turn controls the entry of cells from late G1 to S phase and subsequent S phase arrest. In this regard, we assessed the gene expression of the specific S phase cellular markers p21 and cyclin E1, given the results of cell cycle evaluation. Our results are perfectly in line with what we expected, mirroring initial outcomes from cell cycle analysis. Notably, both NAR and HSP increased p21 expression at different times and concentrations, as well as lowered that of cyclin E1. Interestingly, their combination, also in this case, proved to be effective and comparable to the double of the concentration of the single flavonoids. Therefore, we presumed that THP-1 cell cycle blockage by NAR or HSP may be mediated by p53 activity, but we do not exclude that other p53-independent signaling pathways, closely related to SIRT2, may be involved in the S arrest, as previously suggested [48]. Finally, to understand how NAR and HSP interact with SIRT2, we investigated the putative binding mechanism of these two flavanones by computational techniques. SIRT2 structure was fully deciphered in 2001 by Finnin and co-workers [49] who revealed that SIRT2 possesses a catalytic core domain with NAD+-binding capacity as well as N- and C-terminal extensions. The catalytic core contains a variant of the Rossmann fold and a small domain consisting of helical and zinc-binding modules. The abovementioned domains are separated by a large lipophilic area that represents a wide groove containing an active site where deacetylation of substrates occurs. The catalytic groove accommodates various inhibitors from natural and synthetic sources [29,50]. The X-ray crystal structure of the 2-(4,6-dimethyl-pyrimidin-2-ylsulfanyl)-N-(5-naphthalen-1-ylmethyl-thiazol-2-yl)-acetamide (SirReal2) in complex with SIRT2 (PDB code 4RMG) [25] revealed that the inhibitor occupies the “selectivity pocket” shaped by Ile93, Ala135, Leu138, Pro140, Phe143, Leu206, and Ile213. In this hydrophobic pocket, the dimethylmercaptopyrimidine moiety of SirReal2 establishes π-π stacking interactions with Tyr139 and Phe190. Moreover, the naphthyl group of SirReal2 protrudes into the acetyl-lysine channel that is considered the substrate binding site comprising various hydrophobic residues (Phe131, Leu134, Ile 169, Ile232, Val233, Phe234). In detail, the inhibitor engages π-T-shaped interactions with Phe131 and Phe234.
Hematological malignancies continue to represent a significant challenge, being frequently depicted as incurable diseases. Therefore, urgent development of novel therapeutic agents is needed to overcome the failure of standard therapies and to improve the patients’ survival rate. Some plant-derived products, such as flavonoids, have gained great interest due to their pharmacological potential. Among these, we demonstrated that NAR and HSP, the two most common flavanones of Citrus fruits, can exert anti-leukemic effects on the human leukemic monocytic cell line THP-1 through growth inhibition via the arrest of cell cycle progression. Interestingly, these mechanisms appear to be linked to the inhibition of SIRT2, a known proliferation marker in relapsing AML, which makes flavonoids attractive candidates for the management of this pathology. In addition, we showed that NAR and HSP can reciprocally either enhance their antiproliferative effects, thus providing further new possibilities for studying the combined anti-leukemic activity of natural products. | true | true | true |
|
PMC9599039 | Michela Relucenti,Federica Francescangeli,Maria Laura De Angelis,Vito D’Andrea,Selenia Miglietta,Orlando Donfrancesco,Xiaobo Li,Rui Chen,Ann Zeuner,Giuseppe Familiari | A Different Exosome Secretion Pattern Characterizes Patient-Derived Colorectal Cancer Multicellular Spheroids and Their Mouse Xenografts | 29-09-2022 | colorectal cancer,exosomes,transmission electron microscopy,scanning electron microscopy,spheroid,xenograft | Simple Summary Exosomes have a role in tumorigenesis and metastatic dissemination, their material content and size being associated with poor prognosis of colorectal cancer (CRC). Our work aims to investigate their secretion patterns in CRC stem cells in patient-derived multicellular tumor spheroids (MTSs) and their mouse xenografts, to unveil possible differences in terms of exosome amount, size, and secretion site between in vitro and in vivo models. Our results show that MTSs’ exosome secretion pattern depends on their structural complexity: few-layer spheroids show a lesser exosome secretion, limited to the apical domain of cancer cells; secretion increases in multilayered spheroids and is visible from apical and basolateral cancer cells domains. In xenograft models, exosome secretion occurs from all cancer cell domains, and it is quantitatively greater than that observed in spheroids. The influence of the surrounding environment of non-tumor cells may account for the difference in exosome secretion patterns between spheroids and xenografts. Abstract Up-to-date in vitro and in vivo preclinical models expressing the patient-specific cancer lineage responsible for CRC and its metastatic behavior and responsiveness to therapy are needed. Exosomes’ role in tumorigenesis and the metastatic process was demonstrated, and the material content and size of the exosomes are associated with a poor prognosis of CRC. Exosomes are generally imagined after their recovery from blood serum as isolated entities, and our work aims to investigate them “in situ” in their native environment by scanning and transmission electron microscopy to understand their secretion modalities. We studied CRC stem cells in patient-derived multicellular tumor spheroids (MTSs) and in their mouse xenograft to find possible differences in terms of exosome amount, size, and secretion site between in vitro and in vivo models. We observed that MTSs’ exosome secretion patterns depend on their structural complexity: few-layer MTSs show a lesser exosome secretion, limited to the apical domain of cancer cells, secretion increases in multilayered MTSs, and it develops from apical and basolateral cancer cells domains. In xenograft models, exosome secretion occurs from all cancer cell domains, and it is quantitatively greater than that observed in MTSs. This difference in exosome secretion pattern between MTSs and xenografts may be due to the influence of surrounding non-tumor cells. | A Different Exosome Secretion Pattern Characterizes Patient-Derived Colorectal Cancer Multicellular Spheroids and Their Mouse Xenografts
Exosomes have a role in tumorigenesis and metastatic dissemination, their material content and size being associated with poor prognosis of colorectal cancer (CRC). Our work aims to investigate their secretion patterns in CRC stem cells in patient-derived multicellular tumor spheroids (MTSs) and their mouse xenografts, to unveil possible differences in terms of exosome amount, size, and secretion site between in vitro and in vivo models. Our results show that MTSs’ exosome secretion pattern depends on their structural complexity: few-layer spheroids show a lesser exosome secretion, limited to the apical domain of cancer cells; secretion increases in multilayered spheroids and is visible from apical and basolateral cancer cells domains. In xenograft models, exosome secretion occurs from all cancer cell domains, and it is quantitatively greater than that observed in spheroids. The influence of the surrounding environment of non-tumor cells may account for the difference in exosome secretion patterns between spheroids and xenografts.
Up-to-date in vitro and in vivo preclinical models expressing the patient-specific cancer lineage responsible for CRC and its metastatic behavior and responsiveness to therapy are needed. Exosomes’ role in tumorigenesis and the metastatic process was demonstrated, and the material content and size of the exosomes are associated with a poor prognosis of CRC. Exosomes are generally imagined after their recovery from blood serum as isolated entities, and our work aims to investigate them “in situ” in their native environment by scanning and transmission electron microscopy to understand their secretion modalities. We studied CRC stem cells in patient-derived multicellular tumor spheroids (MTSs) and in their mouse xenograft to find possible differences in terms of exosome amount, size, and secretion site between in vitro and in vivo models. We observed that MTSs’ exosome secretion patterns depend on their structural complexity: few-layer MTSs show a lesser exosome secretion, limited to the apical domain of cancer cells, secretion increases in multilayered MTSs, and it develops from apical and basolateral cancer cells domains. In xenograft models, exosome secretion occurs from all cancer cell domains, and it is quantitatively greater than that observed in MTSs. This difference in exosome secretion pattern between MTSs and xenografts may be due to the influence of surrounding non-tumor cells.
Epidemiological data on colorectal cancer (CRC) incidence and mortality (according to the World Health Organization GLOBOCAN database) show that CRC is the third most commonly diagnosed cancer in males and the second in females around the world, even if marked differences in rates exist among countries [1]. Stand-alone surgery is generally curative in 40% of patients with CRC stages 1 or 2, and 5-year survival rates reach 90% [2]. This approach is not sufficient for the management of advanced stages and metastatic CRC, which represent about 30% of cases at the time of diagnosis [3]. For this kind of patient, chemotherapy, radiotherapy, and their combination are used, but a marked heterogeneity in patients’ clinical responses exists [4,5], accounting for poor survival rates [6]. One strategy to reduce mortality in CRC is, on the one hand, to succeed in finding biomarkers for early detection and, on the other, the development of personalized therapies to treat patients in the more advanced stages of the disease. The oncology community is moving towards patient-tailored cancer therapy, taking into account the unique molecular profile of each patient’s cancer. This innovative approach is improving responses to therapy [7,8], and the development of up-to-date in vitro and in vivo preclinical models expressing the patient-specific cancer lineage and genetic diversity is needed to understand some fundamental aspects of patient-specific genetic alterations, not only in CRC arising but moreover in its metastatic behavior and responsiveness to therapy. In the last two decades, MTSs cultures have been developed from CRC and other tumors and are now considered reliable preclinical in vitro models of cancer [9]. CRC MTSs consist of 3D cultures of primary cells derived from surgical specimens, reproducing patient-specific genetic expression profiles and heterogeneity [10]. At the same time, the role of exosomes as specific biomarkers in CRC prediction and screening is emerging [11]. Exosomes are nano-sized vesicles (30–120 nm), in their single membrane, express a high and cancer-specific glycosylation profile [12] and are the carrier for various lipids, proteins, DNA fragments, and several RNA species as mRNAs, microRNAs (miRNAs), long noncoding RNAs (lncRNAs) as well as small interfering RNAs (siRNAs) [13,14,15,16,17]. The potential role of exosomes in tumorigenesis and the metastatic process was demonstrated [18,19,20]. Exosomes’ role in cancer progression and metastasis is that of carriers, which actively transfer bioactive molecules between cancer cells and different cell types in the nearby and distant microenvironments. The effect of such intercellular cross-talk explicates by changing multiple cellular and biological functions in recipient cells [21,22,23]. Moreover, the poor prognosis of cancer is associated with the material content and the size of the exosomes rather than the frequency of blood circulating exosomes. Exosomes are generally imaged after their recovery from blood serum by drop-casting as isolated entities, and we aim to look at them in their native environment. We focused on an exosome secretion pattern study “in situ,” observed by scanning and transmission electron microscopy patient-derived MTSs and their mouse xenograft, to find possible differences in terms of exosome amount, size, and secretion site between in vitro and in vivo models.
Patient A 63 aged year male underwent CRC surgery for cancer removal under the standards of the ethics committee on human experimentation of the National Institute of Health (Istituto Superiore di Sanità) authorization no.CE5ISS 09/282, as reported in [24]. Cancer biopsies management (immediately after recovery) Samples were washed 2–3 times in cold saline and transferred in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific, Carlsbad, CA, USA, https://www.thermofisher.com, accessed on 27 May 2022), mixed with 3% penicillin-streptomycin-amphotericin B solution (Lonza Group, Walkersville, MD, USA, http://www.lonza.com, accessed on 27 May 2022). Biopsies dissociation procedure Samples were washed 3–4 times in phosphate-buffered saline (PBS), sectioned into small fragments (0.5 × 0.5 mm), and incubated in DMEM (Thermo Fisher Scientific) with 1.5 mg/mL collagenase type II (Thermo Fisher Scientific) and 20 g/mL DNAse (Roche Diagnostics, Indianapolis, IN, USA, https://usdiagnostics.roche.com, accessed on 27 May 2022) for 1 h at 37 °C, under shaking. Cell culture Resuspensions of pellets containing cells, cell clusters, and tissue fragments were cultured in CSC medium supplemented with 10 mm nicotinamide, 1 µm y-27632 (both from Sigma-Aldrich, St. Louis, MO, USA, http://www.sigmaaldrich.com, accessed on 27 May 2022), 20 ng/mL human EGF and 10 ng/mL human basic fibroblast growth factor (both from Peprotech, London, UK, https://www.peprotech.com, accessed on 27 May 2022). For further detail, see [10,24].
Animal procedures were performed according to the Italian National animal experimentation guidelines (D.L.116/92) and upon approval of the experimental protocol by the Italian Ministry of Health’s Animal Experimentation Committee. Animals Four- to 6-week-old female NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice (The Jackson Laboratory, Bar Harbor, ME, USA, https://www.jax.org, accessed on 27 May 2022). For CSC validation, 5 × 105 cells were injected subcutaneously in the flank of 3 replicate mice in 100 µL 1:1 PBS/Matrigel (BD, Franklin Lakes, NJ, USA, http://www.bd.com, accessed on 27 May 2022). In all the validated CSCs, xenografts were detectable within 3–5 weeks in at least 2/3 mice. Xenograft extraction and treatment Palpable xenografts were extracted, and samples were then formalin-fixed and paraffin-embedded. A pathologist evaluated hematoxylin and eosin-stained sections to compare xenograft histology with that of the tumor of human origin.
Primary fixation (immediately upon recovery) MTSs were fixed in a solution of 2.5% glutaraldehyde in Phosphate buffer solution 0.1 M, pH 7.4 at 4 °C for 48 h. Washing: samples were rinsed overnight in Phosphate buffer solution 0.1 M, pH 7.4 at 4 °C. Post-fixation A solution of osmium tetroxide (OsO4) at 1.33% in H2O (Agar Scientific, Stansted, UK) was used to submerge tissue fragments for 2 h. Washing: phosphate buffer solution 0.1 M, pH 7.4 for 20 min (10 + 10 min) was used to remove (OsO4) residuals [25,26,27]. Dehydration procedure Ascending alcohol series (30, 70, 95, and 100% v/v) solutions were used. Critical point drying procedure (Emitech K850, Emitech, Corato, Italy). Samples were mounted on aluminum stubs using carbon tape. Sputter coating procedure. With platinum (Emitech K 550 sputter coater, Emitech, Corato, Italy operating conditions: 15 mA, for 3 min). Observation Hitachi SU 4000 Field emission scanning electron microscope under high vacuum at 20 kV. Digital image acquisition system: DISS5 Digital Image Scanning System (Point Electronic, Halle (Saale), Germany).
Primary fixation (immediately upon recovery) MTSs and xenograft biopsies were fixed in a solution of 2.5% glutaraldehyde in phosphate buffer 0.1 M, pH 7.4 at 4 C for 48 h. Washing: samples were rinsed overnight in Phosphate buffer solution 0.1 M, pH 7.4 at 4 °C. Post-fixation Samples were then post-fixed in a solution of OsO4 1.33% in H2O (Agar Scientific, Stansted, UK) for 2 h. Washing: phosphate buffer solution 0.1 M, pH 7.4 for 20 min (10 min + 10 min) was used to remove (OsO4) residuals. Dehydration procedure Ascending alcohol series (30, 70, 95, and 100% v/v) solutions were used. Substitution procedure Propylene oxide was used (BDH Italia, Milan, Italy), 2 steps of 20 min each. Embedding procedure In a mixture of 50:50 propylene oxide and epoxy resin Agar 100 (SIC, Rome, Italy) overnight at 25 °C (under a chemical fume hood). Finally, samples were embedded in fresh epoxy resin Agar 100 (Agar scientific, Agar Scientific Ltd., Stansted, Essex, UK) and put on a stove at 60 °C for 48 h [28,29,30]. Sectioning procedure: Semithin sections (1 m thick) were collected on glass slides, stained blue by methylene blue, to perform light microscopy observations by a Zeiss Axioskop-40 (Carl Zeiss, Oberkochen, Germany) equipped with Axiovision image acquisition software. Ultrathin sections for TEM observations were cut using an ultramicrotome (Leica EM UC6, Vienna, Austria). Ultrathin sections were collected on 100-mesh copper grids (Assing, Rome, Italy). Staining was performed using Uranyless© solution and lead citrate 3% solution (Electron Microscopy Science, 1560 Industry Road, Hatfield, PA, USA). Imaging procedure: Observation under a transmission electron microscope (Carl Zeiss EM10, Thornwood, NY, USA) set with an accelerating voltage of 60 kV. Digital image acquisition system: CCD digital camera (AMT CCD, Deben UK Ltd., Suffolk, UK).
Exosomes and MVBs diameters (N = 200 for each group: spheroid apical, spheroid basolateral, xenograft apical, xenograft basolateral, MVB Apical, MVB basolateral) were measured on transmission electron microscopy digital images using open source Fiji software [31] and Hitachi 3D Map (Digital Surf, Besancon, France) [32]. Data were statistically analyzed, and summary statistics, t-tests, and ANOVAs with Bonferroni correction were performed, and data were plotted in histograms. All procedures were performed using Med Calc Statistical software (MedCalc Software 20.009 version Ltd., Acacialaan 22, 8400 Ostend, Belgium).
Our morphological investigation started with the analysis of the MTSs’ three-dimensional morphology by scanning electron microscopy; we then proceeded to their ultrastructural characterization; by transmission electron microscopy, highlighting aspects related to the secretion of exosomes. Finally, an ultrastructural analysis of the xenograft was performed to compare the different experimental models and highlight similarities and differences in exosome secretion. Data on the size of exosomes and MVBs were then finally statistically analyzed.
Observation of the outer morphology of MTSs by scanning electron microscopy showed that their outer surface had a variable appearance. Some MTSs appeared as compact entities with smooth surfaces or sparse, shallow furrows (Figure 1A). No exosome secretion was observed from the cells forming the outer surface. Other MTSs showed some shallow surface grooves (Figure 1B), corresponding to the boundaries of the underlying cells; even in this case, no exosome secretion by the outermost cells was observed. Still, other MTSs exhibited on their outer surface deep and numerous grooves (Figure 1C), with one or more cells protruding from the surface of the spheroid itself. The cells of the outermost layer exhibit blebs and microvilli, but no exosome secretion was observed.
Observing the internal morphology of the MTSs, at first by light microscopy on semi-thin sections, then by transmission electron microscopy on ultrathin sections, we identified three different types of MTSs, hereafter defined as A, B, and C. MTSs of group A were characterized as bi-layered or three-layered structures, containing pseudocyst-like structures (resembling a colonic gland; Figure 2A). Cells lining the pseudocyst lumen tightly adhere to each other and project microvilli on their apical surface. In the pseudocyst lumen, no exosome secretion was visible. The cells of the outermost layers appeared loosely adhered to each other and were separated by large intercellular spaces (Figure 2B,C) in which cells’ membrane extroversions extended (Figure 2C,D). These threadlike extroversions were sometimes short, other times were longer and convoluted, intertwining with those of adjacent cells. In the intercellular spaces, no exosome secretion was visible. Group B MTSs were structured in three to five cell layers, and the innermost cells were arranged to form a pseudocyst-like structure (Figure 3A). Cells lining the lumen of the pseudocyst tightly adhered to each other and presented microvilli on the apical surface and secrete (Figure 3B). Cells in the outermost layers appeared more adherent to each other than in the same cells of group A MTSs, but intercellular spaces are still present (Figure 3C). These cells also project finger-like membrane eversions into the intercellular spaces, but we did not observe the secretion of exosomes into these spaces (Figure 3C) or towards the external surface of the spheroid. Group C MTSs have been identified as multilayered structures whose innermost cells arrange to form a pseudocyst-like structure (Figure 4A,B). The cells bordering the lumen of the pseudocyst were tightly adhered to each other, possessed microvilli on the apical surface, and secreted a large number of exosomes (Figure 4B) and entire MVBs. The cells of the outermost layers were also in contact with each other (Figure 4B), few intercellular spaces were present, and their lumen was occupied by digitiform eversions of cell membranes. In this group of MTSs, the same cells with microvilli that excrete exosomes into the lumen of the pseudocyst can also secrete exosomes from the membrane of the lateral domain, and the secretion of exosomes was observed from the middle layers cells’ membrane, which spills exosomes into the intercellular spaces (Figure 4C). No exosome secretion was observed by cells’ membrane of outermost layers towards the external surface of the spheroid, i.e., directly in the culture medium.
The tumor resulting from cancer-derived MTSs xenograft in immunodeficient mice (for simplicity, we will call this the xenograft) is a tissue, a more complex entity with different characteristics than MTSs, which are a 3D cultured cell mass. Simply consider the presence of necrotic and hypoxic areas, blood vessels, nerves, and fibroblasts (Figure 5A); altogether, those factors create a different extracellular environment that is lacking in the MTSs culture system. Xenograft contains cells arranged often in pseudocysts resembling crypts and glands of the colonic tract (Figure 5A,B) with scarce stroma in between. No goblet or enteroendocrine cells were visible as in MTSs from whom the xenograft originated or in the patient’s cancer. Those pseudocysts contained cells with a columnar shape similar to enterocytes, as well as more oval cells with large oval and indented nuclei, which did not open into the gland lumen and mitotic figures. Cell nuclei were dysmorphic, with large nucleoli and heterochromatin aggregates along the inner aspect of the nuclear membrane. To study the secretion patterns of exosomes in xenografts, images obtained by transmission electron microscopy observation of the samples were analyzed. The TEM images shown in Figure 6 demonstrate the presence of secretion activity in the apical domain of cells with microvilli surrounding the lumen of the pseudocyst. This secretion activity is intense and develops homogeneously along the apical surface. The intensity of secretion is caused by the presence of numerous MVBs (Figure 6A–C), aligned in rows perpendicular to the apical surface of the cell (Figure 6A–C). The MVBs release their exosome content at the base of the microvilli (Figure 6C,D), and MVBs are often entirely secreted. Phenomena of exosome secretion were also observed from the basolateral domain of xenograft cells facing the tissue interstitium (Figure 7). In particular, cells with microvilli were observed secreting exosomes into the lumen of pseudocyst from their apical domain but also towards the tissue interstitium from the basolateral domain (Figure 7A,B). This secretory activity is less intense than that along the apical surface. This lower intensity is caused by the fact that there are far fewer MVBs pouring their contents into the intercellular space, and these MVBs are scattered and not organized in parallel rows as they were in the apical domain. Secretory activity of exosomes was also observed from deeper cells that do not face into the lumen of the pseudocyst and that lack polarization in the apical, lateral, and basal domains (Figure 7C,D). No entire MVB secretion was observed from the basolateral cells’ domain. The different patterns of secretion in MTSs and xenografts at both the apical and basolateral membranes, secretion in MTSs and xenograft is shown in Figure 8. In MTSs samples, few and sparse MVBs are aligned in one horizontal row just beneath the apical plasmatic membrane. In xenografts, MVBs are aligned in several vertical columns perpendicular to the apical plasmatic membrane (Figure 8A,B; see also Figure 6). Few MVBs approach the basolateral membrane in MTSs, while groups of generally four or five MVBs are visible near the basolateral membrane in the xenografts (Figure 8C,D). To provide a quantification of different secretion amounts in MTSs and xenografts, we counted the number of MVBs in epithelial cells lining pseudocystic structures (200 cells from Type C MTSs and 200 cells from xenografts). Summary statistics and t-test results are illustrated in Table 1 and Figure 9. MBV secretion from the apical domain of spheroid cells is five times higher than basolateral secretion. MBV secretion from the apical domain of xenograft cells is 10 times higher than basolateral secretion. MBV secretion from the apical domain of xenograft cells is massive compared to secretion from the apical domain of spheroid cells, in a ratio of 13:1. MBVs secretion from the apical domain of xenograft cells is higher than secretion from the apical domain of spheroid cells, in a ratio of 6:1.
Following the observation of MTSs’ and xenografts’ different secretory modalities, we focused our analysis on the morphology and size of exosomes and MVBs. Looking at the appearance of exosomes secreted from the apical domain (in both MTSs and xenograft), we noted that they presented long filaments projecting radially from the outer aspect of the membrane. These molecules partly intertwine at their initial part, proximal to the membrane, and form a “crown” around the outer surface of the exosome. These filaments consist of glycoproteins and glycolipids of exosome membrane, highlighted by the use of tannic acid in the sample preparation process for electron microscopy (Figure 10A,C). Exosomes secreted from the basolateral domain (Figure 10B,D) also showed glycoproteins and glycolipids on their membrane, although in smaller amounts than in exosomes secreted from the apical domain (Figure 10C,D). Exosomes secreted in basal and inflammatory conditions from human epithelial cells had a size range of 30–90 nm [33,34]. We measured the size of exosomes secreted from the different domains in the same sample (spheroid apical vs. spheroid basolateral and xenotransplant apical vs. xenotransplant basolateral), and data were statistically analyzed (Table 2, Figure 11). As can be seen from Figure 10C, the exosomes observed in a tissue sample included in resin and cross-sectioned appear not to have all the same size, as it is when vesicles extracted by centrifugation from culture medium or serum are observed at TEM by drop-casting. In the latter, the real diameter of the vesicles corresponds to the average value of the observed diameters. In our case, the correct interpretation of the real size of the vesicles must be made, taking into account that, since these are spherical dissected structures, the real diameter of the vesicle will correspond not to the average value of the observed diameters but will correspond to their maximum value (Figure 11). The distribution values show that values in the first three columns are about two times higher concerning the last column (70–77 nm), being the value of maximum diameter (sphere equator in Figure 11E) the lesser frequent for the values in the others classes, that recur at least two times. To verify if some difference exists in exosome diameter values in different samples (MTS vs. xenograft) and different secretion sites (apical vs. basolateral), an ANOVA test with Bonferroni correction was performed (Figure 12). The size of MVBs secreted from the apical domain of spheroid and xenograft was measured and compared. MVBs secreted by the spheroid basolateral domain were scarce (being present only in type C MTSs) concerning that of xenotransplant, and no comparison between these two groups was made. Results of statistical data analysis are reported above in Table 3 and Figure 13. To verify if some difference exists between MVB diameter value in MTSs vs. xenografts, a t-test was performed and results are presented in Figure 14.
Cancer stem cells are a tumor subpopulation capable of self-renewal and are crucial for survival, proliferation, drug resistance, metastasis, and tumor recurrence [35]. We recently generated a molecularly characterized biobank of colorectal CSC-enriched lines that represent a priceless resource available for in vitro studies and for the development of CSC-based murine models that faithfully reproduce the molecular and histological features of the primary tumor [10]. This primary tridimensional cell culture of tumor-derived MTSs retains the genetic heterogeneity of the original patient tumor and displays CSC’s ability to dynamically switch between CSC and non-CSC states [35,36]. CRC spheroid cultures also reproduce drug sensitivity profiles of parental tumors, thus representing an excellent preclinical model to investigate the efficacy of new anticancer therapies [10,11]. While the use of MTSs for drug testing has been the object of intense studies, the biological and structural features of MTSs are less explored. We have previously investigated the ultrastructural features of CRC MTSs, highlighting an increased presence of stem-like cells in MTSs as compared to tumor xenografts [24]. In this study, we investigated the presence and localization of exosomes in CRC MTSs and xenografts. Exosomes play a crucial role in mediating cell-to-cell communication between CSCs, non-stem cancer cells, and other cells in the tumor microenvironment (TME), regulating processes such as tumor progression, metastasis, drug resistance, EMT, and immune evasion [37]. The results of our ultrastructural study show that spheroids have exosome secretion patterns that depend on their structural complexity. Precisely, spheroids formed by a few layers of cells (Group A), with little adhesion between them, do not show exosomes production; when the number of cell layers increases and the degree of cell adhesion of the spheroid increases (Group B), secretion of exosomes into the lumen of pseudocysts is observed; until we have, in spheroids with several cell layers and a high degree of intercellular adhesion (Group C), secretion of exosomes and multivesicular bodies both from the apical domain of cells surrounding the lumen of the pseudocysts, both from the basolateral domain of the same cells and also from cells that do not face the lumen. As reported in [38], MTSs from different CRC cell lines organize into three main types: loose, tight, and compact, and this different organization is related to different adhesion molecules expression. Loose MTSs express integrin-mediated interactions that are subsequently substituted by N-cadherin and, finally, E-cadherin interactions. In the study of [33,34,39], the different molecular profiles of apical vs. basolateral exosome secretion were demonstrated. Our ultrastructural results (from a patient-derived CRC cell line) correlate for the first time the molecular data from the literature with the ultrastructural imaging of the different loose, tight, and compact MTS organizations, showing its relation with different exosome secretion patterns and amounts from both the apical and basolateral surfaces. This means that exosome production can be targeted in different pathways [40] to slow its progression. Lower exosome production will result in the lowering of cancer-promoting effects triggered by exosome-carried molecules. In xenograft models (from a patient-derived MTSs xenograft), the secretion of exosomes occurs from all domains of the tumor cells and is quantitatively greater than that observed in spheroids. The massive presence of MVBs that release their contents into the lumen of the pseudocysts or in the tissue interstitium was observed. Our ultrastructural results show for the first time the different arrangements in rows and columns (different patterns) and the difference in the amount of MBV secretion between spheroids and xenografts, xenografts being the source of a massive exosome and MVB production. Our findings suggest that targeting the pathways of MVB formation and release could be another way to slow down cancer progression and translate our findings into clinical applications. This difference in exosome secretion pattern and amount between MTSs and xenografts may be possibly due to the influence of surrounding non-tumor cells, as it has been shown that exosomes are key mediators of the communication between tumor cells and the tumor microenvironment [41]. Statistical analysis conducted on measurements of exosome diameter shows that exosomes secreted from the different MTSs domains have the same size (about 70 nm), which is then also the same in those produced by xenograft cells. According to [42], the size of our observed exosomes corresponds to that of a “small exosome” (Exo-S), ranging from 60 to 80 nm. Exo-S are particularly rich in Flotillin 1, flotillin 2, tweety family member 3, tetraspanin 14, and ESCRT-I subunit VPS37B. The data in [42] indicate that exosome size, in addition to their specific cargo, may influence metastatic patterning and the systemic effects of cancer.
Our morphological data show that structural complexity influences exosome secretion of MTSs, in both intensity and pattern, if compared with xenograft models. Our observations add new knowledge to the ultrastructural features of CRC MTSs and xenografts. Future studies may define the mechanistic basis of different exosome secretion patterns in the two model systems. | true | true | true |
|
PMC9599176 | Helen M. Wise,Adam Harris,Nisha Kriplani,Adam Schofield,Helen Caldwell,Mark J. Arends,Ian M. Overton,Nick R. Leslie | PTEN Protein Phosphatase Activity Is Not Required for Tumour Suppression in the Mouse Prostate | 19-10-2022 | tumour suppressor,prostate cancer,PTEN,phosphatase,PI 3-Kinase | Loss PTEN function is one of the most common events driving aggressive prostate cancers and biochemically, PTEN is a lipid phosphatase which opposes the activation of the oncogenic PI3K-AKT signalling network. However, PTEN also has additional potential mechanisms of action, including protein phosphatase activity. Using a mutant enzyme, PTEN Y138L, which selectively lacks protein phosphatase activity, we characterised genetically modified mice lacking either the full function of PTEN in the prostate gland or only lacking protein phosphatase activity. The phenotypes of mice carrying a single allele of either wild-type Pten or PtenY138L in the prostate were similar, with common prostatic intraepithelial neoplasia (PIN) and similar gene expression profiles. However, the latter group, lacking PTEN protein phosphatase activity additionally showed lymphocyte infiltration around PIN and an increased immune cell gene expression signature. Prostate adenocarcinoma, elevated proliferation and AKT activation were only frequently observed when PTEN was fully deleted. We also identify a common gene expression signature of PTEN loss conserved in other studies (including Nkx3.1, Tnf and Cd44). We provide further insight into tumour development in the prostate driven by loss of PTEN function and show that PTEN protein phosphatase activity is not required for tumour suppression. | PTEN Protein Phosphatase Activity Is Not Required for Tumour Suppression in the Mouse Prostate
Loss PTEN function is one of the most common events driving aggressive prostate cancers and biochemically, PTEN is a lipid phosphatase which opposes the activation of the oncogenic PI3K-AKT signalling network. However, PTEN also has additional potential mechanisms of action, including protein phosphatase activity. Using a mutant enzyme, PTEN Y138L, which selectively lacks protein phosphatase activity, we characterised genetically modified mice lacking either the full function of PTEN in the prostate gland or only lacking protein phosphatase activity. The phenotypes of mice carrying a single allele of either wild-type Pten or PtenY138L in the prostate were similar, with common prostatic intraepithelial neoplasia (PIN) and similar gene expression profiles. However, the latter group, lacking PTEN protein phosphatase activity additionally showed lymphocyte infiltration around PIN and an increased immune cell gene expression signature. Prostate adenocarcinoma, elevated proliferation and AKT activation were only frequently observed when PTEN was fully deleted. We also identify a common gene expression signature of PTEN loss conserved in other studies (including Nkx3.1, Tnf and Cd44). We provide further insight into tumour development in the prostate driven by loss of PTEN function and show that PTEN protein phosphatase activity is not required for tumour suppression.
Loss of function of the PTEN tumour suppressor is among the most frequently observed genetic events that drive prostate cancer [1,2]. The most common mechanism of loss is deletion of a single PTEN gene copy, which is usually associated with loss of PTEN protein expression [2]. Notably, the occurrence of PTEN loss is higher in metastatic disease (seen in 40–60% of these cases) than in primary tumours (10–40%) [3,4,5] and associates with poor prognosis [6,7,8,9,10]. It has therefore been proposed that PTEN status could be used to distinguish between indolent and progressive prostate cancer and accordingly, PTEN is a component of several biomarker signatures which appear to have some power to identify aggressive disease [11,12,13,14,15,16,17,18,19]. PTEN is a core component of the class I phosphoinositide 3-Kinase (PI3K) signalling network, acting as a lipid phosphatase to dephosphorylate the PI3K products, phosphatidylinositol 3,4,5-trisphosphate (PIP3) and probably also phosphatidylinositol 3,4-bisphosphate (PI(3,4)P2) [20,21]. PI3K and its product lipids play conserved roles regulating metabolism, and promoting the growth, proliferation and survival of many cell types and additionally influencing cell polarity in a more lineage specific manner [21,22,23]. These downstream consequences of PI3K activity are mediated by a large diverse group of direct phosphoinositide lipid-binding proteins, amongst which a dominant role is played in many processes by the AKT group of protein serine/threonine kinases [23,24,25]. Accordingly, there have been intense efforts to develop small molecule inhibitors of PI3K, AKT and the downstream kinase mTOR as drugs to treat cancers including prostate cancer. The success of these agents in clinical trials has been disappointing [26], but notably the AKT inhibitor Ipatasertib has recently been shown to increase progression free survival in a phase III trial in prostate cancer patients displaying PTEN loss [27]. Heterozygous Pten+/− mice succumb to a broad range of tumours [28] and genetic deletion of Pten specifically from the developed prostate gland causes rapid high grade prostatic intraepithelial neoplasia (PIN) and later invasive prostate carcinoma, with kinetics which seem to depend on genetic background [29,30]. While there is compelling evidence that the lipid phosphatase activity of PTEN, functionally opposing PI3K, is its dominant mechanism of tumour suppression [31,32,33], PTEN also has other potential mechanisms which may contribute [34,35]. These include phosphatase-independent functions in the nucleus and elsewhere [35,36]. PTEN also has weak but robust phosphatase activity in vitro against protein and phosphopeptide substrates, with highest activity against acidic phosphotyrosine substrates [37], and a number of roles and potential substrates for this protein phosphatase activity have been proposed [38,39,40,41,42,43,44,45]. However, confident determination of the substrates of protein phosphatases is challenging and to date a clear picture is yet to emerge regarding the significance of these proposed substrates in PTEN function. To test the significance of the protein phosphatase activity of PTEN, we have previously engineered a PTEN mutant, PTEN Y138L, which retains activity against lipid substrates yet lacks the normal activity of PTEN against phosphopeptides [46]. PTEN Y138L retains the ability to suppress AKT phosphorylation in cultured cells yet unlike the wild-type enzyme it fails to inhibit glioma cell invasion [47] or control epithelial 3D lumen formation [48]. Notable in these studies, the activity of PTEN Y138L in these cell-based assays could be rescued by mutation of Thr366 but did require lipid phosphatase activity. This indicates that, at least in these assays, the only requirement for the protein phosphatase activity of PTEN is the autodephosphorylation of this residue in the PTEN C-terminus [47,48]. Here, we have generated organ-specific knock-in mice expressing PTEN Y138L to test the requirement for the protein phosphatase activity of PTEN for tumour suppression in the prostate.
Mice: The constitutive endogenous gene knock-in PtenY138L mouse line was developed as a service by Taconic-Artemis (Cologne, Germany) by homologous recombination in C57BL/6 NTac ES cells and has been studied for spontaneous tumour formation in heterozygosity (Priyanka Tibarewal, Laura Spinelli and Nick Leslie unpublished data). These mice had been back-crossed to C57BL/6J mice for at least 10 generations. The conditional Pten allele [49,50] and the prostate-specific PB4-Cre [51] mouse lines have been previously described. The PB4-Cre line was provided by the NCI Frederick Mouse Repository (Frederick, MD, USA). During breeding, the PB4-Cre allele was only allowed to pass through male germline [52]. Mice were maintained in randomly assigned cages in the University of Edinburgh Biological Research Facility following all facility guidelines. All experiments were approved by the University of Edinburgh Ethical Review Committee and performed in accordance with the UK Animals (Scientific Procedures) Act 1986 and following UK Home Office guidance. Mouse PCR genotyping used template DNA released from ear punches with Microzone microLYSIS reagent (Microzone, Stourbridge, UK) and Kapa TAQ (Merck/Sigma, Darmstadt, Germany) following manufacturers protocols. The genotyping primers used were: CreF—CCATCTGCCACCAGCCAG; CreR—TCGCCATCTTCCAGCAGG; PtenY138LF—ATGGAAAGGAGTAAATGGATGG; PtenY138LR—GGAGTAAAAGCAGGAGAATTGG; PtenFloxF—GCCTTACCTAGTAAAGCAAG; PtenFloxR—GGCAAAGAATCTTGGTGTTAC. To compare spontaneous tumorigenesis, mice were maintained until tumour formation was evident by palpation, other signs of ill health or 600 days of age in tumour-free animals. At sacrifice, the prostate, pelvic lymph nodes and visceral organs were fixed for histology and IHC. Any large tumours were divided for both formalin fixation and immediate appropriate lysis for RNA and protein analysis. Antibodies and Immunohistochemistry: The antibodies used for Immunohistochemistry (IHC) were: Phospho-S473 AKT Antibody #9271 from Cell Signaling Technology (also used for immunoblotting); PTEN antibody Clone 138G6 #9559 from Cell Signalling Technology; Androgen Receptor (AR) from DAKO; PKM2 antibody from R&D systems (AF7244); RAC1-GTP mAb from NewEastBio #26903. IHC followed Leica BOND Protocols as follows. Androgen Receptor IHC used Leica’s defined ‘Mouse IHC’ protocol; epitope retrieval (ER) was achieved using solution ER1 for 20 min and a 1/100 primary antibody dilution was used. IHC for Ki67 used the ‘Mouse IHC’ protocol with an extended 1h antibody incubation, ER1 for 20 min and undiluted primary antibody. RAC1 IHC used the ‘Mouse on Mouse’ protocol, ER1 for 20 min and a 1/100 primary antibody dilution. P-473AKT IHC followed the ‘Mouse IHC’ protocol, 20 min ER1 and a 1/50 dilution of primary antibody. Caspase 3 IHN followed the ‘Mouse IHC’ protocol, with 20 min ER2 and a 1/50 primary antibody dilution. Cell Culture and Immunoblotting: LNCaP cells were purchased from ATCC (designated LNCaP clone FGC) and cultured in RPMI-1640 with 10% FBS. Cells were transduced with pHR-SIN lentiviruses encoding human PTEN, cells lysed and protein expression and phosphorylation investigated as previously described [47]. Phospho-S240/244 Ribosomal Protein S6 was from Cell Signaling Technology, clone D68F8, #5364. Immunoblotting (but not IHC) for PTEN used an antibody (clone A2B1) from Santa-Cruz Biotechnology. GAPDH immunoblotting used MAB374 antibody from Merck Millipore. Gene expression analysis: Prostates were dissected from 4 mice of each genotype sacrificed at the ages of 6 weeks and 20 weeks of age. RNA was purified from all pooled lobes using an RNAeasy kit (Qiagen, Hilden Germany) and gene expression analysed by hybridisation to Affymetrix GeneChip Mouse ST2.0microarrays as a service by Cambridge Genomic Services (Cambridge, UK). Microarray data processing and functional analysis: The 6-week samples were processed in a single batch, however the 20 week samples contain data from two different batches and a clear batch effect was observed (Supplementary Figure S1). We applied ComBat [53] to mitigate batch effects followed by RMA with quantile normalisation. Analysis with ArrayQualityMetrics [54] eliminated two samples from the six week dataset and two samples from the 20 week dataset. Annotation data was obtained using the AnnotationDbi and clariomsmousehttranscriptcluster.db Bioconductor packages [55]. Empirical Bayes and topTable approaches were used in the limma package [56] to calculate ANOVA comparisons between samples and Benjamini-Hochberg corrected p-values (q-values) [57]. The following contrasts were used to construct the Contrast Matrix for the 6 week and 20 week eBayes linear models: “WW-WF”, ”WW-FF”, ”WW-YF”, ”WF-YF”, ”YF-FF”, ”WF-FF”. We took thresholds of q < 0.05 and fold-change > 1.5 to identify differentially expressed genes (DEGs). We carried out Functional Annotation Clustering [58] on the DEGs, taking an enrichment score threshold of 1.3 to identify significant Annotation Clusters and using detected genes to define the reference (background) annotations. Comparison across public Pten null prostate gene expression data: Six paired gene expression datasets published and archived in GEO from wild-type and Pten null mouse prostates were identified: GSE25140; GSE56470; GSE96545; GSE76822; GSE24691; GSE98493 [12,30,59,60,61,62]. These were analysed using the GEO2R platform to identify gene lists from each study which vary in their expression between wild-type and Pten null prostates with adjusted p < 0.05. Comparison between the lists and our own 6 week and 20-week data then identified sets of genes which were altered in all four young (6–12 weeks) or all four older (15–30 weeks) groups in each case.
To confirm the ability of PTEN Y138L, a mutant which selectively lacks protein phosphatase activity, to regulate AKT in prostate cells in the manner of wild-type PTEN, it was expressed transiently in unselected LNCaP cells using lentiviral vectors. These prostate cancer cells lack PTEN and display elevated phosphorylation of AKT and S6 which is reduced upon expression either of wild-type PTEN or PTEN Y138L but not mutants lacking lipid phosphatase activity, PTEN C124S or PTEN G129E (Figure 1A and Figure S2). To generate mice with this genetic modification specifically in the prostate gland, we used mice expressing the recombinase Cre driven by the prostate-specific PB4 probasin promoter [51]. These were bred with mice carrying a Pten allele flanked by loxP sites [50] and Pten+/Y138L mice expressing the PTEN Y138L mutant from the endogenous Pten locus. Notably, the embryonic lethality observed in homozygous PtenY138L/Y138L mice limits the opportunity to study this homozygous genotype (Priyanka Tibarewal manuscript in preparation). Four PB4-Cre positive Pten genotypes were used experimentally in our study: Pten+/+, Pten+/flox, PtenY138L/flox and Ptenflox/flox (Figure 1B,C). 26 mice of each genotype were studied. Cohorts of 6 mice were sacrificed at 6 weeks of age and at 20 weeks for histology, RNA and protein analysis. Further groups of 14 mice were monitored up until the age of 85 weeks (Figure 1D). In agreement with previous studies [29,30], PIN had formed in Ptenfl/fl prostates by 6 weeks of age and later, regions of locally invasive adenocarcinoma in each of these mice by 20 weeks of age (Figure 2A,B and Figure S3). In contrast, in prostates carrying a functional Pten gene with lipid phosphatase activity, in either PtenY138L/flox or Pten+/flox mice we observed a much slower development of disease. All of these mice had foci of PIN by 20 weeks of age, but almost all mice had not developed evident carcinoma by this time (Figure 2A,B). In accordance with this apparent pathology, immunohistochemistry (IHC) showed increased proliferation, revealed by Ki67 staining, and elevated detection of activated phosphorylated AKT phosphorylation and active GTP loaded RAC1 in prostates lacking PTEN (Ptenflox/flox) but not in those expressing either the wild-type enzyme or the PTEN Y138L mutant protein (Figure 2A,C and Figure 3A). IHC also revealed strong nuclear staining for the androgen receptor (AR) in the majority of cells in the prostates of 6-week-old Ptenflox/flox mice. However, in samples from each of the other genotypes, AR staining was weaker and heterogeneous (Figure 3B). These data analysing progression from PIN to carcinoma, rates of proliferation and increases in AKT phosphorylation and AR protein levels, suggest that only full deletion of PTEN in Ptenflox/flox mice drives these changes. Accordingly, analysis of survival in the mice cohorts found that Ptenflox/flox mice showed significantly shorter overall survival, with all but one of this group being sacrificed with prostate tumours by 65 weeks of age (Figure 4A). In these phenotypic data, Pten+/flox and PtenY138L/flox mice appeared similar, showing only slow development of PIN. This implies that the dominant factor in these observations may be the lipid phosphatase activity of PTEN correlating with the regulation of AKT. However, there were notable differences between the high-grade PIN lesions observed in the Pten+/flox and PtenY138L/flox mice as only the latter cohort showed frequent large regions of lymphocyte infiltration and of more fibroblastic stroma (Figure 4B–E). Lymphocyte infiltration was observed in PIN in all of the PtenY138L/flox mice and stromal changes observed in samples from three of the six mice. To gain deeper insight into the changes occurring before and during the development and progression of PIN in these mice, global gene expression analysis was conducted. RNA was purified from the prostates of 4 mice from each genotype at 6 weeks and 20 weeks of age and analysed using hybridisation microarrays, followed by data quality control, differential expression analysis and functional enrichment analysis. Notably the abundance of Pten mRNA in Ptenflox/flox prostate tissue was 0.16× that found in Pten+/+ mice tissue (t-test p = 0.0022). The number of genes significantly changed in their expression across the four genotypes reflected the observed pathology (Table 1). At 6 weeks of age, gene expression changes were modest (<200 genes significantly different) and the largest numbers were observed in Ptenflox/flox prostates relative to each of the other 3 genotypes. At 20 weeks of age, the greatest number of significant changes (>800 genes upregulated; >500 genes downregulated) were observed between wild-type and Pten null prostates, with intermediate numbers of genes being differentially expressed between wild-type tissue and prostates carrying a single allele encoding either wild-type or Y138L mutant PTEN. However, the differences in gene expression between prostate tissues of these two genotypes each with a single lipid phosphatase active Pten allele (Pten+/flox and PtenY138L/flox) was very modest (<100 genes differentially expressed). To identify functional patterns within the gene expression data, Functional Annotation Clustering was performed, integrating annotation data from multiple databases [58]. The gene expression signatures from the 6-week prostates indicated consistent patterns of functional change within the Pten null prostate tissues relative to each of the other three genotypes and associating with the development of PIN (Supplementary Table S1). Notably, there were only 5 functional clusters upregulated in the PtenY138L/flox tissue relative to the Pten+/flox tissue, one of which is designated “Inflammatory response/Immunity” and contains genes associated with both innate and adaptive immune systems. This appears consistent with the immune cell infiltration observed during histology (Supplementary Table S1 and Figure 4E). This functional assignation is associated with significantly increased expression in the PtenY138L/flox tissue of eight genes previously linked to immune cell function: Naip5, Thbs1, Elf3, Prkcq, Anxa1, Pglyrp1, Reg3g and Ltf. The availability of gene expression data from several previous studies provided the opportunity to use these to validate consistent changes in gene expression caused by the full deletion of Pten from the murine prostate. Therefore, we compared our data (Supplementary Tables S2 and S3) with those from six published tissue specific knock-out studies which also compared gene expression in Pten+/+ and Pten−/− prostates [12,30,59,60,61,62]. We used three studies from young mice (6–12 weeks) and three from older mice (15–30 weeks) which were compared to our data from 6 weeks and 20 weeks, respectively. This identified 43 genes which were differentially expressed between wild-type and Pten-null prostates in all four studies of young 6–12-week-old mice and 82 genes differentially expressed in prostate tissue in all of the four independent datasets from older mice between 15 and 30 weeks of age (Supplementary Tables S4 and S5). Notably, these consistently observed changes in gene expression between wild-type and Pten null prostates included orthologs of many genes functionally associated with human prostate cancer, such as NKX3.1 [63], TNF [64], CD44 [65], KLF5 [66], ASNS [67], CXCL16 [68], IL1RN [69], LY6A/SCA1 [70], GATA3 [71], TNS1 [72] and STAT1 [73], and encoding a number of reported relevant biomarkers including ANXA2 [74], KRT19 [75], SDC1 [76], RHOU [77], SEPT9 [78], ANPEP [79], TSPAN8 [80], COL4A6 [81] and ATF6 [82] (Supplementary Table S5). The signal for a microarray probe (18746) specific for the oncogenic M2 isoform of pyruvate kinase was increased in both Ptenflox/flox and PtenY138L/flox samples relative to Pten+/+ and Pten+/flox prostates. To confirm this finding, immunohistochemistry specific for the Pyruvate Kinase M2 (PKM2) isoform showed elevated levels in Ptenflox/flox prostate tissue of 6-week-old mice relative to other genotypes, particularly in areas of PIN (Figure 5). A similar pattern was also observed for the Keratin 19 gene (Krt19) which was elevated at the mRNA level in both Ptenflox/flox and PtenY138L/flox samples, but with IHC showing increased expression of KRT19 protein only in Ptenflox/flox samples especially areas of PIN, relative to other genotypes (Figure 5). Notably, KRT19 has been recently described to protect cancer cells from the immune system, acting in concert with transglutaminase 2 (TGM2) [83] which was also found to be elevated at the mRNA level in Ptenflox/flox prostates of 20-week-old mice (Supplementary Table S3).
We have bred colonies of mice lacking one or both copies of the Pten gene specifically in the prostate as well as mice with a single Pten gene copy encoding a mutant enzyme with lipid phosphatase activity but not protein phosphatase activity. A key new conclusion of this work is that adenocarcinoma development is driven by full loss of PTEN function in the prostate in a manner that is not observed when PTEN function is partially reduced or protein phosphatase activity is selectively lost. This correlation provides further support for the functional connection between canonical PI3K-AKT signalling and neoplastic prostate pathology. The observation of PIN in prostates with a single copy of either wild-type Pten or PtenY138L but adenocarcinoma in Pten null prostates is consistent with previous observations of the dose dependency of tumour suppression in Pten mutant mice [29,30]. It is also consistent with the observation that PTEN Y138L and the wild-type enzyme display very similar phosphatase activity against PIP3 and suppression of AKT in cell-based assays (Refs. [46,47] and Figure 1A). Accordingly, the apparent lack of significance of the protein phosphatase activity of PTEN in prostate tumour supprssion supports, and provides insight into, existing efforts to treat prostate cancers with small molecule inhibitors of PI3K, AKT and mTOR. The results of tens of clinical trials with inhibitors of PI3K, AKT and mTOR in prostate cancer patients have been disappointing with significant adverse effects and little or no efficacy [26,84]. A possible contributing factor to these failures has remained our poor understanding of the processes driving and sustaining these cancers, and dysregulation of protein substrates of PTEN acting independently of PI3K in cells lacking PTEN has been proposed [38,39,40,41,85]. In contrast, our data show that the protein phosphatase activity of PTEN is not required for tumour suppression in the prostate and provide support for efforts to optimise the clinical use inhibitors of AKT signalling in combination therapy [27]. Our data are in accord with the observed phenotypes of mice expressing the stable PTEN G129E mutant allele lacking lipid phosphatase activity but retaining protein phosphatase activity. These mice display a worse tumour phenotype than mice carrying a full null allele and argue strongly against strong independent tumour suppressor functions for the protein phosphatase activity of PTEN [32,33]. Notably, recent studies of mammary tumour formation in these mice provide a possible explanation for these findings. They identify the glucocorticoid receptor as a functional target of PTEN protein phosphatase activity, but acting in opposition to its lipid phosphatase activity as loss of PTEN protein phosphatase activity promoted cell death [45]. A recognised weakness of our study, resulting from the use of the available constitutive knock-in PtenY138L allele, is the expression of this allele throughout the PtenY138L/flox animals. Although no pathology or phenotype has been identified in PtenY138L/+ animals at the relevant ages, it is possible that the phenotypes we observe in the prostates of PB4-Cre PtenY138L/flox mice may be influenced by systemic changes or the expression of PTEN Y138L lacking protein phosphatase activity in, for example, immune cells. We also recognise that the prostate pathology observed in PB4-Cre Ptenflox/flox mice cannot model the genetic and phenotypic diversity of human prostate cancer. Therefore, although we see no evidence for it, we cannot exclude a role for the protein phosphatase activity of PTEN in contributing to tumour suppression in other subtypes of prostate cancer driven by other mechanisms. Immunohistochemistry for the Androgen Receptor (AR) in the prostates of 6-week-old mice surprisingly showed an upregulation of AR in Pten null animals, with a nuclear localisation in most cells throughout the gland, even in areas which appeared morphologically normal. Previous evidence shows that elevation of PI3K-AKT signalling by mechanisms including the loss of PTEN leads to the inhibition of AR-induced gene expression [60,86] and accordingly our analysis of gene expression shows reduced expression of the androgen induced genes Gnmt, Nkx3.1 and Pbsn in 6 week-old prostate tissue lacking PTEN. Therefore, increased AR protein levels and nuclear localisation in these animals seems unexpected and may indicate further complexity in the interactions between the AR and PI3K regulatory systems. No significant effects on the abundance of the AR transcript itself were observed in any of our data, implicating effects at the level of the AR protein. It is noted that AR should bind and activate the Probasin promoter CRE transgene in these animals, but its presence in all genotypes seems to make this an unlikely explanation for these data. 120 genes were identified which were changed in their expression in Ptenflox/flox prostates relative to wild-type in all gene expression studies of either young 6–12-week-old mice or older 15–30 week old mice, but only five genes were identified which were downregulated in all 8 samples across both age brackets. One of these five most consistently regulated genes was Nkx3.1, which encodes a key transcriptional regulator and tumour suppressor in prostate cancer. It has been recognised that loss of PTEN reduces NKX3.1 expression, but previous research has focused on regulation at the protein level, rather than effects which reduce Nkx3.1 mRNA levels [63]. These multiple studies identifying distinct mechanisms by which loss of PTEN also reduces NKX3.1 levels implies the existence of a functionally significant conserved link between the two proteins. In broader terms, these data showing the consistency with which gene expression effects are observed in these genetically modified mouse models of prostate cancer also highlights the very limited understanding of how, and with what effects, these many gene expression changes are driven by loss of the PTEN tumour suppressor. The gene expression studies showed one of the functional clusters upregulated in the PtenY138L/flox tissue relative to the Pten+/flox tissue was designated “Inflammatory response/Immunity”, and was consistent with the histologically observed increase in lymphocyte and other immune cell infiltrations around the PIN lesions in these mice, indicating association with loss of PTEN protein phosphatase activity. It seems relevant that pan-immune-inflammation correlates with poor prognosis in CRPC, yet most prostate cancers display low levels of lymphocyte infiltration and respond poorly to immunotherapy [87,88], it seems significant that loss of the protein phosphatase activity of PTEN appears to correlate selectively with immune cell infiltration. | true | true | true |
|
PMC9599242 | Hongjiao Li,Chenglian Xie,Yurong Lu,Kaijing Chang,Feng Guan,Xiang Li | Exosomal miR92a Promotes Cytarabine Resistance in Myelodysplastic Syndromes by Activating Wnt/β-catenin Signal Pathway | 09-10-2022 | MDS/AML,exosomal miR92a,Ara-C resistance,PTEN,Wnt/β-catenin | Cytarabine (Ara-C) has been one of the frontline therapies for clonal hematopoietic stem cell disorders, such as myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML), but Ara-C resistance often occurs and leads to treatment failure. Exosomal microRNAs (miRNAs, miRs) as small noncoding RNA that play important roles in post-transcriptional gene regulation, can be delivered into recipient cells by exosomes and regulate target genes’ expression. miR92a has been reported to be dysregulated in many cancers, including MDS and AML. However, the effects of exosomal miR92a in hematologic malignancies have not been fully investigated. In this study, qualitative analysis showed the significantly enhanced expression of exosomal miR92a in MDS/AML plasma. Subsequent functional assays indicated that exosomal miR92a can be transported and downregulate PTEN in recipient cells and, furthermore, activate the Wnt/β-catenin signaling pathway and interfere with the Ara-C resistance of receipt MDS/AML cells in vitro and in vivo. Altogether, our findings offer novel insights into plasma exosomal miR92a participating in Ara-C resistance in MDS/AML and we propose miR92a as a potential therapeutic target for MDS/AML. | Exosomal miR92a Promotes Cytarabine Resistance in Myelodysplastic Syndromes by Activating Wnt/β-catenin Signal Pathway
Cytarabine (Ara-C) has been one of the frontline therapies for clonal hematopoietic stem cell disorders, such as myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML), but Ara-C resistance often occurs and leads to treatment failure. Exosomal microRNAs (miRNAs, miRs) as small noncoding RNA that play important roles in post-transcriptional gene regulation, can be delivered into recipient cells by exosomes and regulate target genes’ expression. miR92a has been reported to be dysregulated in many cancers, including MDS and AML. However, the effects of exosomal miR92a in hematologic malignancies have not been fully investigated. In this study, qualitative analysis showed the significantly enhanced expression of exosomal miR92a in MDS/AML plasma. Subsequent functional assays indicated that exosomal miR92a can be transported and downregulate PTEN in recipient cells and, furthermore, activate the Wnt/β-catenin signaling pathway and interfere with the Ara-C resistance of receipt MDS/AML cells in vitro and in vivo. Altogether, our findings offer novel insights into plasma exosomal miR92a participating in Ara-C resistance in MDS/AML and we propose miR92a as a potential therapeutic target for MDS/AML.
Myelodysplastic syndromes (MDS) comprise a heterogeneous group of myeloid neoplasms, which are characterized by peripheral blood cytopenia, hematopoietic cell dysplasia, and a variable disease course. About 30% of MDS patients are at high risk of transformation to secondary acute myeloid leukemia (AML) [1]. As one of the common chemotherapy drugs used for MDS and AML therapy, cytarabine (Ara-C) is used clinically in combination with anthracyclines to achieve the best therapeutic effect [2,3,4]. Ara-C enters the cell and is converted into a therapeutically active triphosphate metabolite, Ara-CTP, which enters the nucleus and inhibits DNA synthesis, in turn triggering apoptosis and exerting antileukemic effects [5]. However, most of the MDS/AML patients will develop Ara-C resistance, and not respond to subsequent therapy. Therefore, understanding the molecular mechanisms that contribute to the emergence of Ara-C resistance is required for improving therapeutic outcomes. Accumulating evidence supports the idea that the continuous cross-talk between cancer cell and local/distant environments is required for effective tumor growth and systemic dissemination, having an important role in the development of drug resistance [6,7]. Exosomes are nano-sized membrane-covered structures (with a diameter of 30–150 nm), originated from the endosomal pathway and secreted via exocytosis into extracellular space [8,9]. Exosomes can mediate intercellular communication between donor and recipient cells by enriching and delivering nucleic acids and proteins [10]. Especially, exosomes act as vehicles for exchange of microRNAs (miRNAs) between heterogeneous populations of tumor cells, generating a transmitted drug resistance [11,12]. miRNAs are a class of small noncoding RNAs, which regulate gene expression by binding to the 3′ untranslated regions (UTR) of target mRNAs [13]. Exosomes can enwrap miRNAs, protect them from degradation, deliver them to recipient cells and modulate tumor immunity and the surrounding microenvironment, further facilitating tumor growth, invasion, metastasis, angiogenesis and drug resistance [14,15,16]. miR92a, located on human chromosome 13q32-33, has been studied in gastric cancer, lung cancer, prostate cancer, liver cancer and thyroid cancer [17,18,19,20]. For example, it suppressed Dickkopf 3 (DKK3) transcription to enhance migration and invasion in colon cancer [21,22]. It also promoted cell proliferation by targeting tumor-suppressor gene F-box and WD-40 domain protein 7 (FBXW7) in nonsmall cell lung cancer [21,22]. These data implied the carcinogenesic role of miR92a. Circulating miR92a in plasma has been thought as novel potential biomarkers for AML [23,24]. However, the function of exosomal miR92a in MDS and AML clone cells was still unclear. In this study, we found that the expression of exosomal miR92a in MDS/AML plasma were higher than that in healthy donor (HD). When recipient cells were treated with exosomes derived from MDS/AML plasma, we found exosomes that expressed high miR92a induced a stronger proliferation ability and Ara-C tolerance of recipient cells.
Firstly, we analyzed the prognosis survival curve of miR92a in pan cancer via the TCGA database. The results showed that higher expressions of miR92a-1 and miR92a-2, two precursors of miR92a, resulted in a poor survival rate depending on Kaplan–Meier analysis (Figure 1A). In addition, miR92a expression in primary blood-derived cancer was higher than that of other sample types (Figure 1B). Next, we evaluated the expression of exosomal miR92a in MDS/AML plasma. The exosomes (Exos) from plasma of healthy donors (HD) and MDS/AML patients were isolated via differential centrifugation and ultracentrifugation (Figure 1C). The purified Exos presented a stable expression of marker proteins (Alix, CD81, CD63 and TGS101) by Western blotting (Figure 1D), and the typical characteristic shapes and sizes by transmission electron microscopy (TEM) and nanosight tracking analysis (NTA) (Figure 1E,F). Next, the qRT–PCR results showed that the expression of exosomal miR92a in MDS/AML plasma was significantly higher than that in HD (Figure 1G). According to the exosomal miR92a level in patient plasma compared to that in HD plasma, Exos derived from MDS/AML patient plasma were divided into two groups, high miR92a-Exos and low miR92a-Exos (Suppl Figure S1A). When SKM1 cells were treated with these two groups of Exos, the miR92a level was significantly upregulated in recipient SKM1 cells (Figure 1H). Meanwhile, high miR92a-Exos and low miR92a-Exos induced resistance to Ara-C in recipient SKM1 cells, and low miR92a-Exos induced SKM1 Ara-C resistance to a relatively lower extent than high miR92a-Exos (Figure 1I). Above all, we speculated that exosomal miR92a is correlated to MDS progression and resistance to Ara-C.
To explore the function of exosomal miR92a, we overexpressed hsa-miR-92a-1 and hsa-miR92a-2 in SKM1, termed as SKM1-miR92a1 and SKM1-miR92a2 (Figure 2A,B). When SKM1-miR92a1/2 cells were treated with 2 μM Ara-C, apoptosis rates were lower than parental SKM1 cells (Supplementary Figure S2A). Next, exosomes from SKM1, SKM1-miR92a1/2 (termed as Exos-SKM1, Exos-miR92a1 and Exos-miR92a2) were isolated by differential centrifugation and characterized by Western blotting, TEM and NTA (Supplementary Figure S2B–E). Flow cytometric results indicated that Exos-SKM1, Exos-miR92a1 and Exos-miR92a2 can be uptaken by recipient SKM1 (Figure 2C). Meanwhile, compared to Exos-SKM1, Exos-miR92a1 and Exos-miR92a2 treatment increased the miR92a level in recipient SKM1 (Figure 2D). Under Ara-C treatment, decreased apoptosis in SKM1 was found after Exos-miR92a1/2 treatment (Figure 2E). Meanwhile, Exos-miR92a1/2-treated SKM1 presented increased proliferation with Ara-C treatment (Figure 2F). Using another MDS cell line (ML1) as recipient cells, similar results of cell apoptosis and proliferation with Ara-C treatment were observed (Supplementary Figure S2F–I).
To determine the mechanism of miR92a in mediating Ara-C resistance, we browsed biological target genes of miR-92a using three online servers (TargetScan, miRDB and TarBase) and screened out five potential target genes, ZEB, FGF2, SMAD4, KLF4 and PTEN (Figure 3A). qRT-PCR analysis showed the decreased expressions of ZEB, FGF2, SMAD4, KLF4 and PTEN in SKM1-miR92a1 and SKM1-miR92a2 cells (Figure 3B). Among these five genes, PTEN expression significantly decreased and it was reported to correlate with drug resistance [15,25]. Therefore, we aimed at PTEN as the potential target of miR92a (Figure 3B). In addition, potential binding sites of miR-92a were identified in the PTEN 3′UTR based on the databases above (Figure 3C). The potential binding capacity of miR92a to PTEN was evaluated by inserting a wild-type (WT) or mutant 3′-UTR sequence of PTEN downstream of the luciferase reporter gene in HEK-293T. The dual-luciferase reporter gene assay showed that the activity of luciferase in miR92a-mimic-treated HEK-293T cells was significantly reduced by inserting a PTEN WT sequence but unaffected by inserting a mutant 3′-UTR PTEN sequence (MUT) (Figure 3D). TCGA database analysis further confirmed a negative relation between miR-92a and PTEN expression in AML (n = 206) and HD (n = 18) (Figure 3E).
It has been found that PTEN is associated with the Wnt/β-catenin pathway in drug resistance [26,27]. The activated Wnt signaling pathway activates gene transcription through transporting β-catenin into the nucleus [28]. We then explored β-catenin activation to reveal potential mechanisms underlying the role of miR-92a in increasing Ara-C resistance. The results showed that the increased expression of β-catenin and target genes, Axin2, c-Myc and CyclinD1, of β-catenin was found in Exos-miR92a1/2-treated SKM1 and ML1 (Figure 4A,B and Figure S3A,B). Furthermore, it showed decreased PTEN and increased β-catenin in the nucleus of SKM1 and ML1 after treatment with Exos-miR92a1/2 (Figure 4B and Figure S3C). Next, the TCGA database indicated increased expression of β-catenin in AML samples (n = 173) compared to HD (n = 70) (Figure 4C). Interestingly, when SKM1 and ML1 were incubated with a selective inhibitor of β-catenin, MSAB [15], the apoptosis of Exos-miR92a1/2-treated recipient cells was increased significantly compared to Exos-SKM1, which revealed that MSAB treatment reversed the Ara-C resistance of SKM1 and ML1 caused by Exos-miR92a1/2 (Figure 4D and Figure S3D). Together, these data indicate that exosomal miR92a mediates the Ara-C resistance of SKM1 via inhibiting PTEN and activating β-catenin.
To detect the effect of exosomal miR92a on Ara-C resistance in vivo, we injected mice with SKM1 cells in combination with Ara-C and Exos-miR92a1/2 (Figure 5A). Flow cytometry results showed that the injection of Exos-miR92a1/2 can significantly increase SKM1 proliferation in vivo after treatment with Ara-C (Figure 5B). In addition, immunohistochemistry (IHC) staining showed increased β-catenin and decreased PTEN in the spleen and bone marrow of mice after injecting Exos-miR92a1/2 (Figure 5C).
Ara-C has been the commonly used therapeutic agent for MDS and AML patients for decades. Considerable progress has been made in the development of new treatments for MDS/AML patients, but drug resistance remains a major clinical problem. Increasing evidence indicates the involvement of exosomes in mediating drug resistance through several mechanisms [26,29,30]. Drug-resistant cancer cells are able to pack the chemotherapeutic agents in exosomes and shuttle anti-cancer drugs out of tumor cells [31], or deliver exosomal cargoes containing miRNA and proteins that later induce drug resistance to recipient cells [32]. For example, exosomal miR21 enhanced cisplatin resistance by suppressing the inflammasome activity of NOD-like receptor thermal protein domain-associated protein 3 (NLRP3) [33]. Exosomal miR155 induces chemoresistance through increasing epithelial–mesenchymal transition (EMT) markers in breast cancer [34]. Previous studies have shown that miR92a plays a crucial role in the development of multiple cancers and the dysregulation of miR92a is reported to have potential as a tumor biomarker [17,35]. However, the function of exosomal miR92a during hematological malignant progression remains unclear. It was reported that hepatocellular carcinoma cell (HCC)-derived exosomal miR92a can promote EMT and convert low-metastatic HCCs into high-metastatic HCCs [18]. Moreover, an elevated expression of exosomal miR92a in plasma is positively correlated with the metastasis of HCC [18]. miR92a-enriched exosomes derived from cancer-associated fibroblasts of colorectal cancer (CRC) can be transferred into CRC cells to promote migration, invasion, metastasis, stemness, and drug resistance [36]. Using the TCGA database, we found that the miR92a level was upregulated in pan cancer and an increased level of miR92a is related to poor cancer prognosis. Later, we demonstrated that the exosomal miR92a level was significantly increased in the plasma of MDS/AML patients. Exosomes are not only the cargoes that contain circulating RNAs, DNAs or proteins, but also reflect the pathological statues of originated cells or tissues. The aberrant exosomal miR92a in MDS/AML also suggests that potential donor cells, such as malignant clonal cells, endothelial cells or macrophages, which generate the dysregulated exosomes, need to be investigated in future studies. Our data indicate both high levels of miR92a-Exos and low levels of miR92a-Exos can induce recipient cells that are resistant to Ara-C, and low levels of miR92a-Exos also induced SKM1 Ara-C resistance to a relatively lower extent than high levels of miR92a-Exos. We speculated that miR92a may be the dominant factor in high miR92a-Exos-induced Ara-C resistance of recipient cells, while some other mechanisms may exist in low miR92a-Exos-mediated Ara-C resistance. miRNAs can post-transcriptionally suppress target mRNA expression, mostly through interaction with 3′ UTR. Using the database and molecular assay, we found that PTEN served as a target of exosomal miR92a and the resulting downregulation of PTEN promoted Ara-C resistance in recipient cells via Wnt/β-catenin pathway. PTEN is a ubiquitously expressed tumor suppressor that is commonly inactivated in human sporadic cancers and it is also major negative regulator of the AKT signaling pathway and Wnt/β-catenin signaling pathway [37,38]. Deficient PTEN suppressed by HOX transcript antisense RNA confers adriamycin resistance in AML [39]. Additionally, PTEN deficiency can activate the AKT pathway to sustain refractory AML status through the enhancement of glycolysis and mitochondrial respiration [40]. A recent study revealed the involvement of PTEN and the Wnt/β-catenin pathway in drug resistance [27], and it has been found that PTEN overexpression can block β-catenin-induced urothelial proliferation and tumorigenesis [41]. In addition, the Wnt/β-catenin pathway has been proved to be required for the development of leukemic stem cells in AML [42] and inactivated Wnt/β-catenin can suppress proliferation and P-gp-mediated multidrug resistance in AML [43]. Taken together, our data indicated that interference with the expression of PTEN and the Wnt/β-catenin signaling pathway may reverse the Ara-C resistance caused by exosomal miR92a. In conclusion, this study revealed that plasma exosomal miR92a is upregulated in MDS/AML and can be transport to receipt cells. It can regulate the Ara-C sensitivity of recipient cells by inhibiting PTEN and activating β-catenin. These results indicate that exosomal miR-92a is a potentially useful target for a more effective chemotherapy of MDS/AML patients.
The plasma of heathy donors and MDS/AML patients was obtained from Shanxi Provincial People’s Hospital. The patients’ information is listed in Supplementary Table S1. Written informed consent was obtained from all patients in accordance with the Declaration of Helsinki. Experiments using human tissues were approved by the Research Ethics Committee of Northwest University.
Exosomes were isolated as described previously [44]. In brief, culture supernatants were collected and subjected to successive centrifugations at 300× g for 10 min, 2000× g for 10 min, 10,000× g for 30 min, and 100,000× g for 70 min at 4 °C. Exosome pellets were rinsed with PBS, collected by ultracentrifugation at 100,000× g for 70 min (Optima XE-100 ultracentrifuge; Beckman Coulter Life Sciences; Indianapolis, IN, USA), resuspended in 100 μL PBS, and stored at −80 °C. Exosomes were labeled using Exo-tracker, as described previously [45]. Exosomes were stained with 10 μM Exo-tracker for 30 min at 37 °C and collected by ultracentrifugation. For the purification of exosomes from AML/MDS patient and HD plasma, plasma was subjected to successive centrifugations at 2000× g for 30 min, 12,000× g for 45 min, and 110,000× g for 2 h at 4 °C. Pellets were resuspended in PBS, filtered (pore size 0.22 μm), collected by ultracentrifugation at 110,000× g for 70 min, and resuspended in 100 μL PBS.
Purified exosomes were applied to carbon-coated 400 mesh grids (Electron Microscopy Sciences; Fort Washington, PA, USA) for 5 min, washed with PBS, and stained with 2% uranyl acetate for 30 s, as described previously [46]. Images were obtained by TEM (model H-7650; Hitachi; Tokyo, Japan) at 80 kV.
Exosomes were loaded into a NanoSight LM10 instrument (Malvern; UK) and particles were tracked for 60 s using the NanoSight nanoparticle tracking analysis software program.
The human hematopoietic cell lines SKM1 and ML1 were cultured in RPMI 1640 medium (Biological Industries, Beit Haemek, Israel) and kidney epithelial cell HEK-293T cells were cultured in DMEM medium (Biological Industries) supplemented with 10% Fetal Bovine Serum (FBS) (Biological Industries) and 1% penicillin/streptomycin (Beyotime Biotechnology, Haimen, Jiangsu, China) at 37 °C in a 5% CO2 atmosphere.
A human precursor of miR92a1 and miR92a2 was cloned from SKM1 cells. The primes are listed in Supplementary Table S2. A precursor of an miR92a1/2 sequence was inserted into the pLVX vector (Invitrogen, Carlsbad, CA, USA). Lentivirus packaging was performed for HEK-293T cells by co-transfecting them with a lentiviral packaging plasmid (psPAX2, pMD2.G, Invitrogen), and expression vectors containing target genes or an empty pLVX vector (Invitrogen) used Lipofectamine 2000 Transfection Reagent (Thermo Fisher Scientific, San Jose, CA, USA) according to the manufacturer’s protocol. An empty vector was used as the negative control. SKM1 was infected using the lentivirus supernatant and selected by 2 μg/mL puromycine (Sigma-Aldrich, St. Louis, MO, USA).
Total RNA was isolated using an RNA pure Tissue & Cell Kit (CW Biotech, Beijing, China), and cDNA was synthesized using a ReverTra Ace qRT-PCR RT Kit (Toyobo, Osaka, Japan) and miRNA cDNA Synthesis Kit, with Poly(A) (Applied Biological Materials, Richmond, BC, Canada) as per the manufacturer’s protocol. Amplification and detection were performed with Power SYBR Green Master Mix (Cwbiotech) and Gentier 48R System (Tianlong Technology, Xi’an, China). The primers are listed in Supplementary Table S2. For quantitative analysis, the relative mRNA levels of β-catenin, SMAD4, ZEB2, KLF4 and PTEN were normalized to GAPDH, and the relative miRNA level of miR92a was normalized to U6. The primers are listed in Supplementary Table S2.
Cells were collected and lysed with lysis buffer (10 μL PMSF in 1 mL RIPA). Protein concentration was determined using a BCA Protein Assay Kit (Beyotime). For Western blotting, proteins (30 μg for total cell proteins; 10 μg for exosomal proteins) were separated by electrophoresis in 10% polyacrylamide gel and transferred onto PVDF membranes. Membranes were blocked with 3% bovine serum albumin (BSA, Sigma-Aldrich, St. Louis, MO, USA), and incubated with antibody against Alix, Calnexin (Santa Cruz Biotech, Santa Cruz, CA, USA), and TSG101, CD63, CD81 (Abcam, Cambridge, UK), β-catenin, tubulin or PTEN (Cell Signaling Technology, Beverly, MA, USA), followed by the addition of HRP-conjugated secondary antibody conjugated with horseradish peroxidase (HRP, Beyotime). Bands were visualized with a chemiluminescence kit and photographed using a bioluminescence imaging system (Tanon 4600, Shanghai, China).
Apoptosis and proliferation were analyzed by flow cytometry (ACEA Biosciences; San Diego, CA, USA). MDS clone cells were treated with 10 μg Exos and 2 μM Ara-C for 48 h. For apoptosis analysis, cells were stained with an Annexin V-PE kit (BD Biosciences, CA, USA). For proliferation analysis, cells were stained with an EdU Alexa Fluor 647 kit (Keygen, Nanjing, China), as per the manufacturer’s protocol.
The miR-92a-3p binding sites were predicted using an online website (http://mirdb.org/ accessed on 5 March 2022). The miR-92a-targeted wild-type PTEN 3′UTR and mutated PTEN 3′UTR were amplified via PCR and inserted into luciferase reporter plasmids psi-Check2, termed psi-Check2-PTEN-WT and psi-Check2-PTEN-mut. Primers are shown in Supplementary Table S2. Next, psi-Check2-PTEN-WT or psi-Check2-PTEN-Mut and miR-92a mimic or the negative control miR-NC were co-transfected into HEK-293T cells. After transfection for 48 h, luciferase activity was determined using the Dual Luciferase Reporter Gene Assay Kit (Beyotime).
Female B-NSG mice (NOD-Prkdc scid IL2rg tm1/Bcgen, NSG) were irradiated with 180 cGy. A total 2 × 106 SKM1 cells were injected into NSG mice through the tail vein. Mice were divided into 3 groups randomly and were injected intravenously with 1.5 mg/kg Exos derived from SKM1, SKM1-miR92a1 and SKM1-miR92a2 cells three times a week. After injection with Exos for 14 days, NSG mice were injected intraperitoneally with Ara-C (80 mg/kg) three times a week. The SKM1 in peripheral blood were stained by antibody against hCD45 and analyzed by flow cytometry after injecting Ara-C 6 times. Mice were humanely sacrificed, and spleen and femur bone were collected and stored for further experiments.
All data were statistically analyzed using GraphPad Prism 5.0 (GraphPad software, San Diego, CA, USA). Data are presented as mean ± SEM. The statistical significance of differences between the means of two groups was evaluated by Student’s t-test. Multiple group comparisons were evaluated by ANOVA with Bonferroni’s post hoc test. | true | true | true |
|
PMC9599605 | 36040022 | Jin-Fei Zhang,Liang-Xing Fang,Man-Xia Chang,Ming Cheng,Hui Zhang,Teng-Fei Long,Qian Li,Xin-lei Lian,Jian Sun,Xiao-Ping Liao,Ya-Hong Liu | A Trade-Off for Maintenance of Multidrug-Resistant IncHI2 Plasmids in Salmonella enterica Serovar Typhimurium through Adaptive Evolution | 30-08-2022 | IncHI2 plasmid,Salmonella,fitness cost,adaptive evolution,plasmid stability,compensatory mutation | ABSTRACT Understanding the fitness costs associated with plasmid carriage is a key to better understanding the mechanisms of plasmid maintenance in bacteria. In the current work, we performed multiple serial passages (63 days, 627.8 generations) to identify the compensatory mechanisms that Salmonella enterica serovar Typhimurium ATCC 14028 utilized to maintain the multidrug-resistant (MDR) IncHI2 plasmid pJXP9 in the presence and absence of antibiotic selection. The plasmid pJXP9 was maintained for hundreds of generations even without drug exposure. Endpoint evolved (the endpoint of evolution) S. Typhi murium bearing evolved plasmids displayed decreased growth lag times and a competitive advantage over ancestral pJXP9 plasmid-carrying ATCC 14028 strains. Genomic and transcriptomic analyses revealed that the fitness costs of carrying pJXP9 were derived from both specific plasmid genes and particularly the MDR regions and conjugation transfer region I and conflicts resulting from chromosome-plasmid gene interactions. Correspondingly, plasmid deletions of these regions could compensate for the fitness cost that was due to the plasmid carriage. The deletion extent and range of large fragments on the evolved plasmids, as well as the trajectory of deletion mutation, were related to the antibiotic treatment conditions. Furthermore, it is also adaptive evolution that chromosomal gene mutations and altered mRNA expression correlated with changed physiological functions of the bacterium, such as decreased flagellar motility, increased oxidative stress, and fumaric acid synthesis but increased Cu resistance in a given niche. Our findings indicated that plasmid maintenance evolves via a plasmid-bacterium adaptative evolutionary process that is a trade-off between vertical and horizontal transmission costs along with associated alterations in host bacterial physiology. IMPORTANCE The current idea that compensatory evolution processes can account for the “plasmid paradox” phenomenon associated with the maintenance of large costly plasmids in host bacteria has attracted much attention. Although many compensatory mutations have been discovered through various plasmid-host bacterial evolution experiments, the basis of the compensatory mechanisms and the nature of the bacteria themselves to address the fitness costs remain unclear. In addition, the genetic backgrounds of plasmids and strains involved in previous research were limited and clinical drug resistance such as the poorly understood compensatory evolution among clinically dominant multidrug-resistant plasmids or clones was rarely considered. The IncHI2 plasmid is widely distributed in Salmonella Typhimurium and plays an important role in the emergence and rapid spread of its multidrug resistance. In this study, the predominant multidrug-resistant IncHI2 plasmid pJXP9 and the standard Salmonella Typhimurium ATCC 14028 bacteria were used for evolution experiments under laboratory conditions. Our findings indicated that plasmid maintenance through experimental evolution of plasmid-host bacteria is a trade-off between increasing plasmid vertical transmission and impairing its horizontal transmission and bacterial physiological phenotypes, in which compensatory mutations and altered chromosomal expression profiles collectively contribute to alleviating plasmid-borne fitness cost. These results provided potential insights into understanding the relationship of coexistence between plasmids encoding antibiotic resistance and their bacterial hosts and provided a clue to the adaptive forces that shaped the evolution of these plasmids within bacteria and to predicting the evolution trajectory of antibiotic resistance. | A Trade-Off for Maintenance of Multidrug-Resistant IncHI2 Plasmids in Salmonella enterica Serovar Typhimurium through Adaptive Evolution
Understanding the fitness costs associated with plasmid carriage is a key to better understanding the mechanisms of plasmid maintenance in bacteria. In the current work, we performed multiple serial passages (63 days, 627.8 generations) to identify the compensatory mechanisms that Salmonella enterica serovar Typhimurium ATCC 14028 utilized to maintain the multidrug-resistant (MDR) IncHI2 plasmid pJXP9 in the presence and absence of antibiotic selection. The plasmid pJXP9 was maintained for hundreds of generations even without drug exposure. Endpoint evolved (the endpoint of evolution) S. Typhimurium bearing evolved plasmids displayed decreased growth lag times and a competitive advantage over ancestral pJXP9 plasmid-carrying ATCC 14028 strains. Genomic and transcriptomic analyses revealed that the fitness costs of carrying pJXP9 were derived from both specific plasmid genes and particularly the MDR regions and conjugation transfer region I and conflicts resulting from chromosome-plasmid gene interactions. Correspondingly, plasmid deletions of these regions could compensate for the fitness cost that was due to the plasmid carriage. The deletion extent and range of large fragments on the evolved plasmids, as well as the trajectory of deletion mutation, were related to the antibiotic treatment conditions. Furthermore, it is also adaptive evolution that chromosomal gene mutations and altered mRNA expression correlated with changed physiological functions of the bacterium, such as decreased flagellar motility, increased oxidative stress, and fumaric acid synthesis but increased Cu resistance in a given niche. Our findings indicated that plasmid maintenance evolves via a plasmid-bacterium adaptative evolutionary process that is a trade-off between vertical and horizontal transmission costs along with associated alterations in host bacterial physiology. IMPORTANCE The current idea that compensatory evolution processes can account for the “plasmid paradox” phenomenon associated with the maintenance of large costly plasmids in host bacteria has attracted much attention. Although many compensatory mutations have been discovered through various plasmid-host bacterial evolution experiments, the basis of the compensatory mechanisms and the nature of the bacteria themselves to address the fitness costs remain unclear. In addition, the genetic backgrounds of plasmids and strains involved in previous research were limited and clinical drug resistance such as the poorly understood compensatory evolution among clinically dominant multidrug-resistant plasmids or clones was rarely considered. The IncHI2 plasmid is widely distributed in Salmonella Typhimurium and plays an important role in the emergence and rapid spread of its multidrug resistance. In this study, the predominant multidrug-resistant IncHI2 plasmid pJXP9 and the standard Salmonella Typhimurium ATCC 14028 bacteria were used for evolution experiments under laboratory conditions. Our findings indicated that plasmid maintenance through experimental evolution of plasmid-host bacteria is a trade-off between increasing plasmid vertical transmission and impairing its horizontal transmission and bacterial physiological phenotypes, in which compensatory mutations and altered chromosomal expression profiles collectively contribute to alleviating plasmid-borne fitness cost. These results provided potential insights into understanding the relationship of coexistence between plasmids encoding antibiotic resistance and their bacterial hosts and provided a clue to the adaptive forces that shaped the evolution of these plasmids within bacteria and to predicting the evolution trajectory of antibiotic resistance.
Antibiotics have been a major accomplishment of modern medicine, but these compounds have suffered a loss of efficacy due to the emergence and dissemination of resistance among bacterial pathogens (1, 2). Therefore, novel strategies are needed to ensure clinical efficacy and to effectively curb the development and spread of bacterial resistance. Bacterial plasmids encode a wide range of phenotypic traits that allow bacteria to adapt to stressors such as the presence of antibiotics and have a key role in bacterial ecology and evolution (3). Interestingly, plasmids are maintained in bacterial populations over the long term, even in the absence of selection for plasmid-encoded traits. This “plasmid paradox” remains poorly understood, since plasmid carriage itself incurs physiological fitness costs to the host bacterium (4). Therefore, understanding the adaptive forces that shape the evolution of these plasmids within a bacterial host may explain the widespread distribution and stable maintenance of plasmids. This information could be used to predict the evolutionary trajectory of antibiotic resistance (5). Salmonella is an important zoonotic pathogen, and currently, multidrug-resistant (MDR) strains that are resistant to fluoroquinolones, third-generation cephalosporins, and even colistin, such as Salmonella enterica serovar Typhimurium and its variants, have emerged (6, 7). Incompatibility HI2 (IncHI2) plasmids are often large (>200 kb) with similar backbone structures and possess two sets of conjugation systems along with numerous antibiotic resistance genes (ARGs) (8, 9). These IncHI2 plasmids are widespread among MDR Enterobacteriaceae and especially S. Typhimurium, even though they incur a high fitness cost for their maintenance (8–10). In this study, we performed 63 serial passages to explore the compensatory mechanisms of adaptive evolution of the MDR IncHI2 plasmid pJXP9 within S. Typhimurium ATCC 14028 in the presence and absence of antibiotics. Our findings suggest that plasmid maintenance through plasmid-host bacterial coadaptation and coevolution is a trade-off but that coevolution of this large IncHI2 plasmid promoted host cell growth and competitiveness. This coevolution thereby increased plasmid vertical transmission by increasing host competitiveness and decreased growth lag times but impaired horizontal transmission of the plasmid and altered host bacterial physiology in a given niche. These results provide insights into understanding the mechanisms that bacteria use to offset the costs of stable plasmid maintenance.
The experimental system we used for this study was a comparison of the cost of maintaining a large IncHI2 plasmid (pJXP9) in S. Typhimurium strain ATCC 14028 with that of its plasmid-free counterpart. Preliminary experiments indicated that carriage of pJXP9 impaired competition and generated a slight growth disadvantage for strain ATCC 14028 (see Fig. S1 in the supplemental material). We extended these experiments and performed serial dilutions of cultures over 63 days in the presence and absence of antibiotics to investigate the stability of the IncHI2 plasmid in endpoint evolved clones/populations (clones/populations from the endpoint of evolution). Using PCR for 240 endpoint evolved clones and quantitative real-time PCR (qPCR) for 12 endpoint evolved populations, we found that IncHI2 plasmid pJXP9 was retained at day 63 with almost no loss when cultured in the presence of ciprofloxacin (CIP) and cefotaxime (CTX) but that plasmid levels were slightly reduced in the presence of colistin (CST) (24% loss using PCR and 7.7% loss using qPCR screening). Interestingly, culture in the absence of antibiotics led to significant plasmid loss (41.7% loss using PCR and 52.3% loss using qPCR screening) (Fig. 1A and B; Fig. S2). These results indicate that carriage of pJXP9 imposed a fitness cost, but the plasmid was maintained for at least hundreds of generations even in the absence of positive selection. To determine the fitness response to plasmid-bacterial host coevolution with or without antibiotic treatment, we directly competed 5 randomly selected endpoint evolved clones from each treatment (total, 20) against strain ATCC 14028::lux+pJXP9, which possessed a luciferase fluorescent marker integrated into the host chromosome. We found an increase in the competitive fitness in all 20 endpoint evolved clones (relative fitness [RF] range, 1.015 to 1.206) except one (RF = 0.7482) (Fig. 1C). These results revealed that adaptive evolution had occurred between the IncHI2 plasmid pJXP9 and the host, ATCC 14028. 10.1128/msystems.00248-22.1 Estimating the fitness effects of pJXP9 plasmid carriage. (A) Growth of ancestral strain ATCC 14028 with the pJXP9 plasmid. (B) Relative fitness of pJXP9 plasmid-carrying versus plasmid-free ancestral ATCC 14028 using competition assays. (C and D) Growth of ancestral ATCC 14028 versus ATCC 14028::lux for plasmid-free (C) and pJXP9-carrying (D) strains. Download FIG S1, TIF file, 0.7 MB. Copyright © 2022 Zhang et al. 2022 Zhang et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. 10.1128/msystems.00248-22.2 Gene abundance analysis by qPCR of clinically important ARGs among endpoint evolved populations. Download FIG S2, TIF file, 1.2 MB. Copyright © 2022 Zhang et al. 2022 Zhang et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. To establish whether the fitness alterations were due to evolution of the plasmid, the bacterial host, or both, the evolved pJXP9 plasmids from evolved hosts were transferred to host ATCC 14028::lux, and 7 transconjugants carrying evolved plasmids were successfully obtained. The competitive fitness measurements for all these transconjugants remained less than those of ATCC 14028+pJXP9 (RF range, 0.88 to 1.01) (Fig. 1D). We further quantified the fitness levels for the 7 evolved clones and corresponding transconjugants using growth curves. Compared to the 7 transconjugants, the corresponding evolved clones displayed significant reductions in growth lag times (47.05 versus 54.30 min) (P < 0.05) (Fig. 1E) but also a decreased growth rate (0.01059 versus 0.01076 per min) (P > 0.05) (Fig. 1F) and a decreased maximum optical density at 600 nm (max OD600) (0.5109 versus 0.5594) (P < 0.05) (Fig. 1G). In contrast, compared to the ancestral host ATCC 14028+pJXP9, the corresponding transconjugants of these evolved plasmids possessed higher growth rates (0.01131 versus 0.01076 per min) (P > 0.05) (Fig. 1F) and max OD600 values (0.5833 versus 0.5594) (P > 0.05) (Fig. 1G) and decreased lag times (53.23 versus 54.30 min) (P > 0.05) (Fig. 1E). Taken together, these findings suggest that evolution of both the plasmid and the host contribute to fitness effect changes in endpoint evolved clones. The competitive advantage and reduced lag times were primarily due to the evolved host, while carriage of the evolved pJXP9 plasmid slightly improved the growth rate and maximum culture density but impaired the competitive fitness.
To determine the contribution of pJXP9 plasmid evolution to cost alleviation, whole-genome sequencing (WGS) data of the pJXP9 plasmids from 20 endpoint evolved clones using 4 different antibiotic conditions were utilized to identify any gene deletions or additions that may have occurred during endpoint evolution experiments (Fig. 2A; Fig. S3). In general, the derived plasmids ranged in size from ~80 to ~244 kb, and the GC content tended to decline (see figure at https://doi.org/10.6084/m9.figshare.20416359). We identified six types of deletions and rearrangements. Type I included 3 clones with deletions of <30-kb fragments. The other five deletion types included the following plasmid regions: MDR region I (MDR I) (n = 3, type II), MDR region (n = 7, type III), MDR I and MDR region II (MDR II) and conjugative transfer region I (n = 4, type IV), MDR I and II and conjugative transfer region I (n = 1, type V), and an ~165-kb region including repHI2 (n = 2, type VI), which was found in two evolved clones, E-CST-S72 and E-CST-S74. Interestingly, deletion sizes were related to antibiotic selection as follows: CST, range of ~20 to ~165 kb; CTX, ~20 to ~50 kb; CIP, ~30 to ~90 kb; and no antibiotic, ~30 to ~79 kb (Fig. 2B). These results indicated a diversity of large fragment deletions for the plasmid MDR and conjugative transfer regions. 10.1128/msystems.00248-22.3 Sequence comparison of evolved plasmid pJXP9 from 10 evolved clones sequenced by Illumina and Nanopore. Sequence comparison of evolved pJXP9 with ancestral pJXP9 as the reference was conducted using the BLAST Ring Image Generator (BRIG) v0.95. The same color series indicates that the plasmids are derived from clones under the same exposure condition. Download FIG S3, TIF file, 2.9 MB. Copyright © 2022 Zhang et al. 2022 Zhang et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
We further determined the relative abundance in evolved IncHI2 plasmid pJXP9 by using the WGS data for 12 endpoint evolved populations under different treatments compared with ancestral plasmid pJXP9 populations. We found that alterations had occurred primarily in the MDR regions between umuC and dcm (~56 kb), including MDR I containing oqxAB between umuC and pJXP9-70 (~33 kb) (D1) and MDR II containing floR, blaCTX-M-14, and fosA3 between pJXP9-70 and dcm (~23 kb) (D2) (Fig. 3A and B). Notably, the relative abundance of MDR I and MDR II depended on the treatments: MDR I was retained at >60% with CIP and >25% with CTX, but gene loss was ~100% with CST and no antibiotic. In contrast, MDR II was retained at 100% with CTX, 65% with CIP, and 25% with no antibiotic and at 15% with CST. Regions (~85 kb, B2 + E + B3) containing mcr-1, ter-like genes, and hypothetical protein genes between dcm and parR largely increased in relative abundance to ~150% but declined to ~70% in region I containing the conjugative transfer region I (C1, ~38 kb) between insJ and umuC and a large hypothetical protein region (B1, ~56 kb) between pJXP9-9 and umuC under CST treatment. Furthermore, relative abundances were slightly decreased to ~89% in the conjugative transfer region I (C1) under CIP treatment. However, the relative abundance was slightly increased to ~120% in the regions (~76 kb, between terY2 and insA/insB) containing a hypothetical protein region (B3, ~23 kb) between terY2 and trhI and the conjugative transfer region II (C2, ~36 kb) between trhI and htdZ in the absence of antibiotics (Fig. S4). 10.1128/msystems.00248-22.4 Relative abundances of genes in plasmid pJXP9 from evolved populations during 63 serial passages under ciprofloxacin and nondrug exposure. (A) Relative abundance of genes in plasmid pJXP9 under ciprofloxacin exposure. E-CIP-1, -14, -28, -42, or -63 denotes populations selected at day 1, 14, 28, 42, or 63 under ciprofloxacin exposure, respectively. (B) Relative abundance of genes in plasmid pJXP9 under nondrug exposure. E-nondrug-1, -14, -28, -42, or -63 denotes populations selected at day 1, 14, 28, 42, or 63 under nondrug exposure, respectively. Download FIG S4, TIF file, 2.6 MB. Copyright © 2022 Zhang et al. 2022 Zhang et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. Additionally, gene relative abundances were largely increased for several insertion sequences, including IS15DI (containing 3 mutated codons compared with IS26) adjacent to oqxA and ISApl1 adjacent to mcr-1 and IS26 adjacent to blaCTX-M14. Intriguingly, the increased extent of gene relative abundance for these insertion sequences largely depended on treatment conditions. For example, the gene relative abundance of IS15DI was increased under CIP and CTX treatments, whereas ISApl1 relative abundance was largely increased in the absence of antibiotics, followed by that under CST, CIP, and CTX treatments. The gene relative abundances for IS26 increased under CTX, followed by that under CIP and no treatment. The gene relative abundance remained almost stable in the other regions or genes on evolved plasmids under different treatments compared with ancestral plasmid pJXP9-carrying populations. Taken together, these results were consistent with the sequence analysis of pJXP9 from endpoint evolved clones in which the MDR and conjugative transfer regions were frequently deleted in evolved plasmids and these deletions depended on antibiotic treatment conditions.
The stability of clinically relevant ARGs in endpoint evolved clones and populations was further investigated using PCR (240 clones), qPCR (12 populations), and WGS (12 populations) analyses (Fig. 3C and D; Fig. S2). We found that the stability of mcr-1, blaCTX-M-14, fosA3, oqxAB, and floR was highly variable and differed between antibiotic treatments. Under CIP treatment, all five tested ARGs remained at high relative abundance, with detection rates of >70% and mcr-1 abundance at ~100%. CTX treatment resulted in high abundance in all tested ARGs but oqxAB, which was retained at ~25%. Unexpectedly, under CST treatment, mcr-1 was completely retained but all of the other four ARGs were almost completely lost at a rate of ~90% and oqxAB was lost at ~100%. Additionally, serial transfer in the absence of antibiotic selection resulted in retention of mcr-1 at ~80%, followed by that of blaCTX-M-14, fosA3, and floR at ~45%. In contrast, oqxAB was almost completely lost (loss rate, ~95%). Antimicrobial susceptibility testing also indicated that resistance phenotypes of the 240 evolved clones were consistent with the presence of oqxAB, mcr-1, blaCTX-M-14, fosA3, and floR (Table S3). Taken together, these results indicate that similar to the MDR regions deletions, the stability of clinically relevant ARGs carried by evolved plasmids were also antibiotic treatment dependent. 10.1128/msystems.00248-22.9 Summary of MIC and PCR data for all 240 evolved clones in evolution experiments. Download Table S3, DOCX file, 0.02 MB. Copyright © 2022 Zhang et al. 2022 Zhang et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
To determine the order of the large fragment deletion mutations in evolved plasmids, we investigated the temporal dynamics of pJXP9 and ATCC 14028 coadaptation evolution under ciprofloxacin and nondrug treatments by WGS using time series-evolved populations (days 1, 14, 28, 42, and 63) (Fig. S4). Under CIP exposure, genes in MDR I (excluding insertion sequences) were almost completely maintained and were stable across 63 serial passages, while those in MDR II (excluding insertion sequences) and C1 decreased to ~70% at day 63. Unexpectedly, the region containing oqxR-oqxA-oqxB and the class I integron harboring aad1 (aminoglycoside resistance)-cmlA (chloramphenicol resistance)-estX (streptothricin acetyltransferase) were completely lost by day 42 and returned to the normal levels found in the ancestral population at day 63. In contrast, passages in the absence of antibiotics generated a gradual decline of MDR II that was slight at days 14 and 28 and more pronounced at days 42 and 63. MDR I also showed a slight decrease at days 14 and 28 and then decreased sharply at day 42 and was completely lost by day 63. Furthermore, the partial conjugative transfer region I (between traJ and umuC) showed slight decreases at days 42 and 63 (Fig. S4). Taken together, for pJXP9 populations evolved under CIP, the class I integration region and oqxR-oqxA-oqxB began to be deleted at least at day 42 and were then followed by deletions of MDR II and C1 by day 63. In contrast, for pJXP9 populations evolved under the absence of drug selection, large fragment deletions first occurred in MDR I and were then followed by deletions in MDR II and the partial conjugative transfer region I. Furthermore, the abundance of IS15DI (adjacent to oqxA), IS26 (adjacent to blaCTX-M14), and ISApl1 (adjacent to mcr-1) tended toward increasing, and the temporal dynamics was also observed under CIP and nondrug treatments.
To analyze whether putative chromosomal modifications mitigate the fitness costs of pJXP9 plasmid carriage, we determined the complete genome sequences of 20 endpoint evolved clones, 12 endpoint evolved populations, and 2 ancestral clones as described above, as well as 8 controls (14,028 evolved clones/population). A total of 81 shared mutated chromosomal genes were identified after excluding mutations also found in the ancestral clones and controls. The mutations identified were common in 13 genes, with frequencies of ≥45% among both multiple evolved clones and populations. These genes were primarily associated with oxidative stress (ahpC, ybgS, STM14_1959, and STM14_2022), DNA repair (umuC and alkB), outer membrane permeability (ompC), osmotic stress (osmY), and sugar transporter (yjiJ), as well as hypothetical proteins (STM14_2712, STM14_3239, STM14_3253, and STM14_4565) (Fig. 4A and B). The gene ahpC encodes an alkyl hydroperoxide reductase subunit belonging to a two-cysteine peroxiredoxin family and is responsible for protecting cells from low concentrations of H2O2 (11). The hypothetical protein YbgS containing two cysteine residues was probably associated with redox reactions (12). STM14_1959 and STM14_2022 encode a putative oxidoreductase and oxidase, respectively. These two genes act together with ahpC and ybgS as antioxidant-related genes, and inactivation of these genes most likely alters the ability of the oxidative stress response. AlkB is an alkylated DNA repair protein (13). DNA polymerase V subunit gene umuC is directly involved in the induction of mutagenesis and associated with the SOS response (14, 15). Mutations in these two DNA repair genes could enhance mismatch repair. The outer membrane porin C protein OmpC allows for ions and hydrophilic solutes to cross the outer membrane, and its mutation could lead to impairment of the structural integrity and change in the permeability of the outer membrane (13). Molecular chaperone OsmY is associated with osmotic shock and the growth state of bacteria (16–19), and mutation inactivation of this gene could impair osmotic stress and result in delayed growth compared with that of wild-type S. Typhi in SPI-2-inducing conditions (17). YjiJ is a putative sugar transporter, and its mutation might be associated with glucose metabolic disorders. To directly test whether loss of gene function plays a role in ameliorating the cost of plasmid carriage, ahpC, osmY, and ybgS were selected, and knockout mutants of these genes were constructed in the ancestral ATCC 14028 background carrying ancestral and evolved plasmid pJXP9. We measured the fitness relative to the wild type with plasmid pJXP9. We found that with ΔahpC, ΔosmY, or ΔybgS, ATCC 14028 mutants carrying ancestral plasmid pJXP9 had higher competitive advantages and slightly lower growth rates than the ancestral strain ATCC 14028+pJXP9 (Fig. 4C and D). Furthermore, ΔahpC, ΔosmY, or ΔybgS mutants carrying evolved pJXP9 also exhibited slightly higher competitive advantage than ancestral ATCC 14028+pJXP9 (Fig. 4E). This indicated that these gene deletions could generally alleviate the cost produced by possession of pJXP9 to some extent. Furthermore, multiple gene mutations were identified in endpoint evolved populations or clones that were under CIP or nondrug exposure (Fig. 4B). Taken together, adaptive coevolution of plasmid pJXP9 within S. Typhimurium ATCC 14028, particularly in the presence of CIP or with no drug, promoted the competitiveness of the evolved bacterial host through compensatory mutations in primarily stress response genes.
Phenotype and genome analyses of evolved populations and clones indicated that the carriage costs of pJXP9 plasmid in Salmonella could be ameliorated through mutations in chromosomal genes associated with the stress response and large fragment deletions in the conjugative transfer regions and multidrug resistance regions in the plasmid. This suggested that there were convergent physiological responses to the carriage of and compensation with this plasmid. To understand these responses and their resolution, we performed RNA sequencing (RNA-seq) on endpoint evolved isolates carrying nearly complete plasmid pJXP9 (3 isolates) and incomplete ones with different deleted regions, from ~33 kb to ~165 kb (4 isolates) (Fig. 5A to G), with or without chromosomal amelioration mutations (Fig. 5I). The pJXP9 plasmid-carrying ancestral clone was chosen as a control. In these 7 evolved isolates, chromosomal gene expression was altered, and the numbers of significantly upregulated and downregulated genes ranged from 185 to 773 and from 169 to 778, respectively (Fig. 5A to G). Furthermore, the number of significantly downregulated genes was slightly higher than that of upregulated genes in 5/7 evolved isolates. The opposite situation was observed in the remaining 2 isolates (Fig. 5D and G). Notably, a set of 16 shared chromosomal differentially expressed genes (DEGs) were identified in all of these 7 evolved clones. These included 5 upregulated genes (scsC, scsD, yegN, STM14_2889, and fxsA) and 5 downregulated genes (ydjN, fliC, dcuB/dcuA, and aspA), as well as 6 genes that were up- or downregulated in different evolved isolates (Fig. 5H). AspA is linked to the tricarboxylic acid (TCA) cycle and aspartate metabolic pathways and catalyzes the conversion of fumarate to l-aspartate (20). dcuA/dcuB encodes common C4-dicarboxylate/aspartate transporters that are active under anaerobiosis (20, 21). The downregulation of these 3 genes (AspA, dcuA and dcuB) indicated that aspartate metabolic pathways might be inhibited. Interestingly, the fimbrial gene fliC, which encodes a structural flagellar protein (22), was downregulated, and this would reduce flagellar motility. YdjN functions as a transporter of S-sulfocysteine, a sulfur-containing intermediate in assimilatory cysteine biosynthesis (23), and its downregulation might impair cystine metabolism and oxidative stress responses (24). scsABCD encode 4 proteins in Salmonella that resemble the disulfide folding machinery of other bacteria, and upregulation of scsBC is linked to adaptation to Cu and H2O2 stress (25). yegN is part of a tripartite efflux system, and its deletion results in decreased growth (26). FxsA overproduction in Escherichia coli inhibits the F plasmid-mediated exclusion of bacteriophage T7 and interacts with the F plasmid-encoded PifA protein to minimize membrane damage (27). Our transcriptome data from evolved plasmid pJXP9, which remained intact from evolved isolates E-CIP-S1, E-CTX-S32, and E-CST-S99, revealed that plasmid-borne genes located on the conjugative transfer regions (C1 and C2) were almost all downregulated except for parAB and parMR. Almost all of the MDR genes were upregulated, with a few exceptions (Fig. S5). Of note, the corresponding resistance phenotypes in these three clones were almost not changed (Table S3). Taken together, the gene expression alterations in sets of shared chromosomal and plasmid-borne genes in the endpoint evolved clones would most likely result in impaired fumaric acid synthesis, flagellar motility, and antioxidant activity, as well as decreased conjugation and bacteriophage-mediated horizontal gene transfer but enhanced resistance against copper. 10.1128/msystems.00248-22.5 Alterations in mRNA abundance in MDR region (A) and transferable regions (B and C) of pJXP9 from evolved clones bearing complete evolved plasmid pJXP9 against ancestral ATCC 14028 carrying pJXP9. Download FIG S5, TIF file, 0.9 MB. Copyright © 2022 Zhang et al. 2022 Zhang et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
IncHI2 plasmids are relatively large (>200 kb), and how they are successfully maintained or persist in the bacterial host is unknown, in particular in S. Typhimurium. We found that the fitness cost imposed on S. Typhimurium ATCC 14028 by IncHI2 plasmid pJXP9 carriage was partially alleviated through coculture of IncHI2 plasmids and S. Typhimurium ATCC 14028. Specifically, experimental evolution was performed for hundreds of generations to achieve potential mutation or fitness-improved information (28). These compensatory mutations included deletions in the plasmid MDR and conjugative transfer regions and specific hypothetical protein regions in evolved plasmids (Fig. 6). Interestingly, these changes were dependent of the presence and type of antibiotic, and similar scenarios of large plasmid MDR region deletions have been reported for evolved plasmid pKP33 in Klebsiella pneumoniae and evolved plasmid pUR2940 in Staphylococcus aureus lineages (5, 29). Indeed, these deletions that we found in the MDR region enhanced the growth rate of transconjugants compared with that of ancestral pJXP9-carrying clones. Furthermore, the evolved isolates E-CIP-S1, E-CTX-S32, and E-CST-S99 carried evolved plasmid pJXP9 without any large fragment deletion, while almost all gene expression from transconjugative regions I and II were downregulated except for that of parAB and parMR (see Fig. S5 in the supplemental material). Conjugation is energetically expensive for host cells, and impairing conjugative transfer probably benefits plasmid maintenance in the bacterial host (30). parAB and parMR are inserted into the transfer gene clusters and capable of supporting partitioning in IncHI2 plasmids (31), and since these genes were not downregulated, their expression most likely facilitated plasmid partitioning and maintenance in the bacterial host. However, unexpectedly, increased expression of MDR genes was observed on evolved plasmid pJXP9 from evolved isolates E-CIP-S1, E-CTX-S32, and E-CST-S99, although MDR region deletions were also common in other evolved clones or evolved populations. This process was most likely linked to an increase in survival ability under antibiotic exposure. Taken together, the conjugation transfer and MDR regions seemed to be one of the primary factors that exerted fitness costs on carriage of IncHI2 plasmids while adaptive plasmid evolution includes deletions that promoted host bacterial growth while retaining the plasmid. We also identified numerous parallel compensatory mutations that appeared in evolved plasmids that included many types of fragment deletions, and interestingly, the extent and range of these deletions on the plasmid in endpoint evolved populations could be linked to culture conditions. Similarly, the differentiated trajectory of deletion mutations in evolved plasmids was also found to be related to selection pressure during the serial evolutionary process under ciprofloxacin and nondrug exposure. But the reasons why gene abundance and deletion mutations occurred between drug exposure groups at specific times remain unclear. One possible explanation is that insertion sequences (ISs) most likely were the arbiters of fragment deletion and recombination (29, 32), and this was reflected in the high gene relative abundance of IS elements in evolved populations. This was especially true for oqxAB, blaCTX-M14, and mcr-1 under corresponding antibiotic treatment conditions. Indeed, ISs contribute to ARG transmission in response to antibiotic exposure and are involved in plasmid-bacterium coadaptation (33–35). Taken together, these results illustrated the dynamics and complexity of plasmid-bacterium coadaptation and coevolution under different antibiotic treatments and time points. The ISs played a significant role in IncHI2 plasmid evolution. Specific parallel compensatory mutations or mutation trajectories were probably due to the directional selection of specific treatments with time, and it made plasmid carriage advantageous. In addition to compensatory deletion mutations in evolved plasmids, bacterial hosts have evolved diverse chromosomal compensatory mutations, such as the global regulatory system genes gacA/gacS, the DNA helicase genes uvrD and xpd/rad3, the nonselective outer membrane porin gene ompF, the efflux pump genes acrAB/acrR, and PFLU4242 of unknown function (36–39). We found a series of chromosomal mutations in stress response genes in evolved clones and, in particular, under CIP exposure or the lack of a drug. We identified deletions in oxidative stress-associated genes, such as the alkyl hydroperoxide reductase gene ahpC and other related but uncharacterized genes (ybgS, STM14_1959, and STM14_2022) and genes for DNA repair (umuC and alkB), outer membrane permeability (ompC), and osmotic stress (osmY). The deletions of ahpC, osmY, and ybgS in the ancestral ATCC 14028 background carrying plasmid pJXP9 improved the competitive capacity of the host. Furthermore, other downregulated genes were independent of antibiotic exposure and were related to bacterial physiological functions, including impaired flagellar motility (fliC), blocked fumaric acid synthesis (dcuB/A and aspA), decreased resistance to oxidative stress (ydjN), inhibited bacteriophage-mediated horizontal gene transfer (fxsA), and enhanced Cu resistance (scsCD). Plasmid acquisition is often associated with reduced bacterial motility and is sometimes due to downregulation of flagellar genes (40). In this study, the swimming zones were significantly inhibited in endpoint evolved strains compared to those of ancestral strains that carried pJXP9 (P < 0.01), whereas the swimming zones in transconjugants carrying evolved plasmids were restored and comparable to those of ancestral strains with ancestral pJXP9 (Fig. S6; see also Text S1 in the supplemental material). These data indicate that impaired flagellar motility was most likely the result of the evolved chromosome. This also implied that potential conflicts between plasmid-bearing and chromosomal genes were also a significant source of fitness cost of carrying pJXP9. As a result, both chromosomal compensatory mutations and altered chromosomal expression profiles influenced bacterial physiology phenotypes to reduce plasmid-carrying costs through coevolution. These indicated that plasmid fitness costs caused by specific genetic conflicts are unlikely to act as a long-term barrier for the persistence of plasmids (36). Furthermore, plasmids can also manipulate chromosomal gene expression (40). Therefore, further studies are required to determine how adaptive coevolution links these bacterial phenotypes to plasmid fitness. 10.1128/msystems.00248-22.6 Comparison of swimming motilities among ancestral and evolved clones and transconjugants by swimming motility experiments. (A) Comparison of swimming motilities between seven picked endpoint evolved clones carrying evolved plasmid pJXP9 and ancestral ATCC 14028 bearing pJXP9. (B) Comparison of swimming motilities between seven picked endpoint evolved clones carrying evolved plasmid pJXP9 and its corresponding transconjugants (ancestral ATCC 14028 carrying evolved plasmid pJXP9). Download FIG S6, TIF file, 0.8 MB. Copyright © 2022 Zhang et al. 2022 Zhang et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. 10.1128/msystems.00248-22.10 Supplemental methods. Download Text S1, DOCX file, 0.03 MB. Copyright © 2022 Zhang et al. 2022 Zhang et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. There are other ecological and evolutionary solutions to the plasmid paradox besides compensatory evolution (41). For example, chromosomal adaptive mutations controlling the global regulatory systems for carbon catabolite repression (CCR) and anaerobic metabolism (ArcAB) reduced the fitness cost of MDR plasmid carriage in specific bacterial niche adaptations (42). Furthermore, bacterial communities that contain multiple species can act as surrogate hosts to maintain a plasmid in the population (42). Moreover, an additional explanation for the maintenance of conjugative plasmids has been proposed, i.e., that plasmid donor cells can also be effective competitors with plasmid-free cells particularly in structured environments (28). Further study is required to determine whether other ecological and evolutionary mechanisms of plasmid stability, such as piggybacking on niche adaptation, can alter the maintenance of IncHI2 plasmids in S. Typhimurium. This study proposes that the source of fitness costs of transferable MDR IncHI2 plasmids depends on plasmid gene expression as well as conflicts between these regions and the chromosome. As a result, compensatory plasmid-borne and chromosomal mutations and altered chromosomal expression profiles collectively contributed to alleviating the fitness costs during adaptative coevolution of the plasmid and the host bacterium. Of note, this adaptive coevolution is also a trade-off between improved growth and competitiveness, with impairments in bacterial physiological processes such as flagellar motility, horizontal gene transfer, cell membrane permeability, fumaric acid synthesis, and resistance to oxidative and osmotic stress. Considering that the specifics of evolutionary dynamics likely vary with the environment and organism, further studies are still required to reveal the nature of adaptive evolution of plasmids and bacteria.
The plasmid pJXP9 (~244 kb) was derived from an E. coli J53 transconjugant of S. Typhimurium JXP9 recovered from a pig in Jiangxi, China, in 2017, possessed a typical IncHI2 plasmid backbone, and carried at least 15 ARGs, including mcr-1, blaCTX-M-14, fosA3, oqxAB, and floR (9). The plasmid pJXP9 was introduced into S. Typhimurium ATCC 14028 by conjugation, and transconjugants were defined as the initial ancestral strain (denoted ATCC 14028+pJXP9) and used for experimental evolution (for details, see Text S1 in the supplemental material). Briefly, cultures were grown in 5 mL Luria Bertani (LB) broth in 50-mL tubes at 37°C and shaken at 180 rpm. Independent selection lines were founded using 12 independent single colonies of ATCC 14028+pJXP9 taken from plate cultures, grown overnight in LB in the absence of antibiotics (nonselective conditions), and split into four exposure groups with biological triplicates: (i) no antibiotic selection, (ii) 2 μg/mL colistin (CST), (iii) 1 μg/mL cefotaxime (CTX), and (iv) 0.015 μg/mL ciprofloxacin (CIP). In parallel, three independent ATCC 14028 colonies were picked up for control treatments and grown in the absence of selection. Cultures were serially diluted 1/1,000 every consecutive day (24 h) for 63 days (63 × log2 1,000) to achieve ~627.8 generations calculated as previously described (28, 43). The final tally for these experiments was 15 endpoint evolved populations (5 exposure scenarios and biological triplicates for each exposure) and a total of 300 endpoint clones (20 clones were selected randomly from each evolved population), which were stored in 30% glycerol at −80°C and used later for phenotyping, whole-genome sequencing (WGS), and RNA sequencing (Fig. 7). Additionally, evolved populations derived from days 1, 14, 28, and 42 that included 3 populations had been passaged in the absence of antibiotics, and 3 populations in the presence of CIP were also stored at −80°C and further deep sequenced (see below).
The stability of pJXP9 and its ARG content were determined at the end of the selection experiment using PCR screening of 20 randomly picked colonies from each tested endpoint population (240 total) using primer sets specific for the plasmid genes repHI2, mcr-1, blaCTX-M-14, oqxAB, fosA3, and floR (Table S1). Plasmid drug resistance was assessed by determining the MICs for 6 antibiotics, including colistin (CST), florfenicol (FFC), cefotaxime (CTX), fosfomycin (FOS), ciprofloxacin (CIP), and nalidixic acid (NAL), in the selected colonies (total, 300 endpoint clones), and the results were interpreted according to the CLSI and veterinary CLSI guidelines (44, 45). MICs for FOS were determined using Mueller-Hinton agar supplemented with 25 μg/mL glucose-6-phosphate. Escherichia coli ATCC 25922 served as the quality control strain. 10.1128/msystems.00248-22.7 PCR primers used in this study. Download Table S1, DOCX file, 0.03 MB. Copyright © 2022 Zhang et al. 2022 Zhang et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. The relative abundances of the 6 plasmid target genes in endpoint populations were determined using quantitative real-time PCR (qPCR) as previously described (2). Briefly, total DNA was extracted from evolved and ancestral pJXP9-bearing populations that were grown in LB broth at 37°C for 10 h. qPCR assays were performed in triplicate using the primers listed in Table S1. The 16S rRNA gene was used as an internal control for DNA quantification. Relative quantification was calculated using the 2−ΔΔCT method (46) [ΔΔCT = (CT, target − CT, control)evolved populations − (CT, target − CT, control)ancestral populations].
Growth curves and competition assays are frequently used methods for estimating plasmid fitness effects (47). Growth curves were measured using triplicate overnight cultures that were diluted to an OD600 of 0.1 in LB broth and distributed in 96-well plates at 200 μL per well. The assay plates were placed into an EnSight multimode plate reader (PerkinElmer, USA) and incubated at 37°C with continuous shaking at 180 rpm, and growth was recorded by measuring the optical density (λ = 600 nm) for a minimum of 10 h. These data were used to extract parameters that served as proxies for bacterial fitness: (i) maximum growth rate (estimating the intrinsic population growth rate), (ii) maximum optical density (max OD600, carrying capacity), and (iii) lag phase duration (lag time, defined as the integrated time lost during adaptation to new conditions compared with an immediate response) (48, 49). Data were processed using the Growthcurver package in R (50). Growth curves were also constructed for ancestral strain ATCC 14028 and ATCC 14028+pJXP9 as well as for endpoint clones bearing evolved plasmid (for details, see Text S1). The relative fitness (RF) of plasmid-carrying versus plasmid-free clones or ancestral plasmid-carrying versus evolved plasmid-carrying clones was estimated using direct in vitro competition assays in triplicate as previously described with slight modifications (51). In brief, growth competition was initiated using a strain cultured for 24 h in LB medium at 37°C and then diluted to an OD600 of 0.1, mixed in a 1:1 ratio, incubated at 37°C for 24 h (day 1), and then diluted to obtain separate colonies when plated on LB agar containing 2 μg/mL CTX or no antibiotic to obtain CFU. Competition assays using ancestral ATCC 14028 versus the ancestral pJXP9-bearing ATCC 14028 strain were performed for 7 consecutive days following the described methods. Competition assays of evolved ATCC 14028 strains bearing the evolved pJXP9 versus ATCC 14028::lux+pJXP9 and of ancestral ATCC 14028 bearing the evolved plasmid versus ATCC 14028::lux+pJXP9 were performed in 1 day. Colonies were screened for epifluorescence using a Leitz Aristoplan microscope and compared to the total plate CFU. The relative fitness was calculated using the equation RF = (log10 S1dt − log10 S1d0)/(log10 S2dt − log10 S2d0), where S1 and S2 represent CFU densities of two tested clones (t = time in days), d0 and dt represent the time when CFU densities of clone S1 or clone S2 were determined respectively (51). An RF of >1 indicated a selective advantage over the control strain, whereas an RF of <1 represented a fitness cost.
Representative evolved clones were sequenced based on MIC and PCR results. Genomic DNA was extracted using a Gentra Puregene bacterial DNA purification kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The DNA was sequenced using an Illumina HiSeq platform (Novogene, China), and sufficient sequencing depth (>100) was obtained for further analysis (1 Gb per clone and 6 Gb per population). The raw data of genomes were filtered by Trimmomatic with -phred 33 (v 0.32), and the clean data were assembled to calculate the estimated genome change using SPAdes (v3.6.2) (-t 30 -k 21,33,55,77,99,127 –careful –phred-offset 33) (52) (Table S2). 10.1128/msystems.00248-22.8 Detailed information for clean data and assembled contigs of 20 evolved clones. Download Table S2, DOCX file, 0.02 MB. Copyright © 2022 Zhang et al. 2022 Zhang et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. High-quality reads were mapped to the plasmid pJXP9 sequence using SOAP aligner/SOAP2 with a maximum alignment error of 5 (53). The relative gene abundances of evolved plasmids (all genes in pJXP9) from evolved populations for transfers at days 1, 14, 28, 42, and 63 were normalized using RPKM (reads per kilobase per million reads) with ancestral pJXP9 as the reference (54). Alignments between evolved plasmids and ancestral plasmids and generation of circular maps were performed using BRIG (v 0.95) (55). Nanopore sequencing was further conducted to obtain the complete nucleotide sequence of plasmid pJXP9 in endpoint evolved clones. For clones sequenced by Nanopore and Illumina, sequence assembly was performed using Unicycler using default parameters for hybrid assembly (v 0.4.8) (56). Plasmid size and GC content alterations for evolved chromosome and plasmid were compared with those of ancestral references using Snapgene (v5.0). The sizes of IncHI2 plasmids were measured from 2 ancestral and 20 evolved strains after linearization of genomic DNA using S1 nuclease (TaKaRa, Dalian, China) by pulsed-field gel electrophoresis (PFGE), followed by Southern blotting hybridization using a digoxigenin-labeled probe specific for repHI2 (Table S1) as previously described (57) (for details, see Text S1).
Sequencing reads of endpoint ATCC 14028+pJXP9 evolved populations and clones were mapped to the ATCC 14028 reference genome (accession no. CP001363 downloaded from https://www.ncbi.nlm.nih.gov/nuccore/CP001363/), and basic variants were called using the CLC Genomics Workbench 10.0 (Qiagen, Hilden, Germany) with default parameters (basic variant detection, similarity fraction of 0.9, minimum coverage of 100, minimum count of 10, and minimum frequency of 10.0%). All chromosomal mutations occurring in >10% of the reads and in at least 10 unique reads were included in the analysis, while those occurring in noncoding regions were excluded (58, 59). Then, the mutations that were present in corresponding endpoint ATCC 14028 evolved populations and clones were filtered out from the above-described mutation results. To better reflect the origin of mutation among populations and clones, we clustered different mutations into the corresponding genes using the following rules: (i) mutations that occurred in >10% in either evolved populations or clones were defined as valid and (ii) cutoffs for gene mutation frequencies among evolved populations and clones were set to acquire the top 25% of the mutated genes. To confirm the impact of chromosomal gene mutation on fitness cost of pJXP9 plasmid carriage, the selected target genes were deleted in ancestral strain ATCC 14028 bearing plasmid pJXP9/evolved plasmid pJXP9 by use of the two-plasmid system pCaspa-pSGKp (based on CRISPR/Cas9-mediated genome editing) CRISPR-Cas9 methods as described in Text S1. The designed 20-nt base-pairing region (N20) of sgRNA for deleting targeted genes and primers for detecting fragment deletion are listed in Table S1. Growth curves and competitive fitness assays were performed among the mutation strains as described above.
In order to analyze the gene expression levels of endpoint evolved clones with evolved plasmid, RNA was extracted from the selected clones cultured in LB broth for 4 h without antibiotics as previously described (60, 61). Bacterial cells were collected by centrifugation at 8,000 × g for 10 min. The library was constructed using an Illumina TruSeq RNA sample prep kit v2 as previously described (62) and sequenced using the Illumina HiSeq 2000 platform. High-quality reads that passed the Illumina quality filter with Trimmomatic -phred 33 were mapped to the S. Typhimurium strain ATCC 14028 genome using the FANSe 2 algorithm with parameters -L55 -E2 -U1 -S10. Genes with ≥10 mapped reads were considered confidently detected genes. Gene expression levels were estimated using RPKM (reads per kilobase per million reads) (54). Gene differential expression analysis was performed as previously described (63). The genes with a <0.001 false discovery rate (FDR) and a change of >2-fold or <0.5-fold were detected as differentially expressed genes (DEGs).
Statistical analyses were performed using Prism 8.0 (GraphPad, San Diego, CA, USA). All data were obtained from at least three biological replicates and are presented as the mean ± standard deviation (SD). Unpaired Student's t test (nonparametric) between two groups and one-way analysis of variance (ANOVA) (nonparametric) among multiple groups with a post hoc test were used to calculate P values. Significance levels are indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001.
Supplementary figures that support the findings in this research were submitted to the figshare database (https://doi.org/10.6084/m9.figshare.20416359).
The population sequencing data and RNA sequencing data reported in this paper have been deposited in the NCBI SRA and GEO databases (BioProject accession no. PRJNA810373 and PRJNA810452 and GEO accession no. GSE197475), respectively. Complete sequences of the evolved strains sequenced by Nanopore and Illumina have been deposited in the GenBank database under accession numbers SAMN26490219 to SAMN26490228. All other data related to this study are available upon request. | true | true | true |
PMC9599699 | Zhaohe Huang,Sitong Liu,Xiaojing Pei,Shujing Li,Yifan He,Yigang Tong,Guoqi Liu | Fluorescence Signal-Readout of CRISPR/Cas Biosensors for Nucleic Acid Detection | 20-09-2022 | CRISPR,fluorescence,nucleic acids detection | The CRISPR/Cas system is now being used extensively in nucleic acid detection applications, particularly after the trans-cleavage activity of several Cas effectors was found. A CRISPR/Cas system combined with multiple signal-readout techniques has been developed for various molecular diagnostics applications. Fluorescence is now a widely utilized dominant read-out technique in CRISPR biosensors. An in-depth understanding of various fluorescence readout types and variables affecting the fluorescence signals can facilitate better experimental designs to effectively improve the analytical performance. There are the following two commonly used types of CRISPR/Cas detection modes: the first is based on binding activity, such as Cas9 and dCas9; the second is based on cleavage activity, such as Cas12a, Cas12b, Cas13, and Cas14. In this review, fluorescence signal-readout strategies from the last 5 years based on the binding activity and cleavage activity of the CRISPR/Cas system with fundamentals and examples are fully discussed. A detailed comparison of the available fluorescent reporter sequences and design principles is summarized. Current challenges and further applications of CRISPR-based detection methods will be discussed according to the most recent developments. | Fluorescence Signal-Readout of CRISPR/Cas Biosensors for Nucleic Acid Detection
The CRISPR/Cas system is now being used extensively in nucleic acid detection applications, particularly after the trans-cleavage activity of several Cas effectors was found. A CRISPR/Cas system combined with multiple signal-readout techniques has been developed for various molecular diagnostics applications. Fluorescence is now a widely utilized dominant read-out technique in CRISPR biosensors. An in-depth understanding of various fluorescence readout types and variables affecting the fluorescence signals can facilitate better experimental designs to effectively improve the analytical performance. There are the following two commonly used types of CRISPR/Cas detection modes: the first is based on binding activity, such as Cas9 and dCas9; the second is based on cleavage activity, such as Cas12a, Cas12b, Cas13, and Cas14. In this review, fluorescence signal-readout strategies from the last 5 years based on the binding activity and cleavage activity of the CRISPR/Cas system with fundamentals and examples are fully discussed. A detailed comparison of the available fluorescent reporter sequences and design principles is summarized. Current challenges and further applications of CRISPR-based detection methods will be discussed according to the most recent developments.
Nucleic acid is the carrier of genetic information and can be used as a biomarker for many disease diagnoses. Most existing nucleic acid assays rely on amplification of nucleic acid, with high cost, low throughput, and complex operations. Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) and CRISPR-associated (Cas) system is an acquired immune system found in most bacteria and all archaea [1,2]. Since its initial discovery, the number of different CRISPR–Cas systems has developed speedily. The CRISPR/Cas system contains two classes (class 1 and class 2) and six subtypes [3]. Cas effectors from class 2, such as type II (Cas9), type V (Cas12), and type VI (Cas13), are commonly used in genome editing because they can efficiently conduct both target identification and cleavage activities with a single protein [4,5,6,7]. It shows great potential in nucleic acid detection, which relies on CRISPR RNA (crRNA) or a single-guide RNA (sgRNA) to direct Cas effector proteins to bind to specific nucleic acid sequences and process them or other nucleic acid sequences, such as binding or cleavage [1,2,8,9]. Cas9, dCas9, Cas12a, Cas13a, and Cas14 are the most often employed Cas effectors and play a vital role in molecular diagnostics applications. By modifying the sequence of crRNA, any target nucleic acid can be easily detected. The CRISPR/Cas system has been developed in collaboration with a range of signal readout methods for diverse target assays due to the precise binding activity and powerful cleavage activity of the CRISPR/Cas system [10,11,12,13]. The Cas9 includes HNH and RuvC-like nuclease domains, which each cleave one target double-strand DNA (dsDNA) via the following two RNA components: mature crRNA and trans-activating crRNA (tracrRNA) [14,15,16]. Furthermore, sgRNA was found to direct Cas9 binding and cleavage by combining the sequences of crRNA and tracrRNA [17,18]. The protospacer adjacent motif (PAM) sequence, which is present in both complementary and non-complementary DNA strands, is the location of the site-specific cleavage, which occurs in three base pairs upstream [5]. dCas9, the catalytically inactive Cas9, lacks cleavage activity compared to Cas9 by suppressing two mutations in the structural domains of RuvC1 and HNH nucleases [15]. The dCas9/sgRNA complex, on the other hand, preserves great specificity for target DNA and has been employed as a nickase in biosensors [19]. Unlike Cas9, Cas12a recognizes target dsDNA using a short T nucleotide-rich PAM sequence [12]. A single RuvC domain is enough to cleave both strands of target DNA using a single crRNA as a guide, producing a PAM-distal dsDNA break with staggered 5’ and 3’ ends [20]. Additionally, Cas12a may also completely disintegrate non-specific single-strand DNA (ssDNA) after binding to a target, a process known as “trans cleavage”. Cas13a, such as Cas12a, possesses trans cleavage ability for non-specific nucleic acids after attaching to target RNA [21,22]. Yet, Cas13a completely cuts non-target RNA after forming the guide-target RNA duplex and employs two conserved HEPN domains to cut target RNA at uracil sites [11]. Furthermore, rather than a PAM sequence, Cas13a detects the target RNA through a 3’ protospacer flanking site (PFS) [23]. Unlike the PAM-proximal seed area required by dsDNA-targeting and the PAM-distal sequence recognized by Cas12a, Cas14’s identification is mediated by interactions around the center of the ssDNA target. Similar to Cas12a, Cas14 detects a target and initiates non-specific ssDNA trans-cleavage activity [10]. The Cas effectors listed above may bind the target nucleic acids with a 20–30 bp long guide RNA (sgRNA or crRNA). Additionally, PAM sequence recognition is essential for Cas effectors to recognize target dsDNA, causing the target dsDNA helix to unwind and enabling the guide RNA to hybridize it [24]. A reliable integrated biosensor should be capable of identifying a target of interest as well as signal transduction for readout. Although CRISPR/Cas-based biosensors have been summarized in some recent reviews. For example, Kaminski et al. summarized various Cas enzymes in clinical diagnostic tests, and Wang et al. reviewed the signal amplification and output of CRISPR/Cas-based biosensing systems [25,26]. The signal readout of CRISPR/Cas has never been classified systematically and discussed comprehensively. In general, CRISPR/Cas biosensors can transduce signals in a variety of ways, such as various optical and electrochemical signals [19,27,28,29,30,31,32,33]. Fluorescence is widely used because of the outstanding advantages in qualitative and quantitative when compared to other methods. In this mini-review, we emphasize the fluorescence signal-readout strategies based on the binding activity and cleavage activity of commonly used CRISPR/Cas systems from fluorescent intensity signals, fluorescent digital signals, fluorescent nanomaterials assisted signals, fluorescent naked eye signals, and others. Figure 1 shows the components of these CRISPR/Cas systems and the identification and cleavage methods, as well as the signal read-out techniques. For each Cas effector, we listed some typical signal transduction categories and main features with representative work (Table 1). The difficulties and challenges currently encountered in applying CRISPR/Cas systems to the assay process will also be discussed.
The fluorescence intensity-based signal readout is easy to encode and provides a simple and effective way for CRISPR biosensors. Most of its analytical signals are obtained based on real-time fluorescence intensity measurements, which describe the average behavior of a large number of molecules over a period of time. Fluorescent probes, such as fluorophores, quantum dots (QDs), nanoclusters, carbon dots, and fluorescent nanomaterials, which display either increased or quenched fluorescence intensity in response to the target objects. The detection methods used by different Cas proteases based on binding and cleavage activity employing fluorescent molecules as signal labels by measurement of fluorescence intensity will be the main topic of this section.
Cas9 and dCas9/sgRNA complexes maintain high specificity for target DNA and have been employed in biosensors and molecular diagnostics. Cas9 is able to recognize and cut the dsDNA. The sgRNA first finds a specific PAM sequence (5′–NGG–3′) in a non-target DNA strand, distant 10–12 nucleotides of the PAM. The target and non-target DNA strands are bound by HNH and RuVC responsible for the cleavage activity, respectively. The deactivated enzyme dCas9 works as a recognition element, specifically binding to the target DNA sequence without cutting it [1,56,57,58]. Metal-organic framework (MOF), which are nonporous materials having structures based on classical coordination bonds between metal cations and electron donors, which has been used for fluorescence quenching/recovery by adsorption/desorption of fluorophore-labeled ssDNA. Sun et al., used the MOF structure (UiO66) to develop an SDA-RCA amplification method with CRISPR/Cas9 for fluorescence detection of E. coli [35]. Following the recognition and cleavage of a one-strand nick of target DNA by two Cas9 sgRNA complexes, strand displacement synthesis expanded at the nick and displaced the original DNA strand. Pre-incorporation of a fluorescent probe bound to UiO66 into the system. When the long ssDNA binds to the probe, it separates them from UiO66, and the quenched fluorescence is recovered. As a result, the recovery fluorescence intensity can be used to quantify the target DNA. There are fluorescent dyes that bind specifically to dsDNA, such as SYBR Green I and ethidium bromide. A common feature of these dyes is that they are not fluorescent unless incorporated into the backbone of dsDNA. As the amount of dsDNA increases during the reaction, the fluorescent signals increase correspondingly. With the SYBR Green I for noncovalent DNA staining, Huang et al. created a CRISPR/Cas9 induced exponential amplification approach (CAS-EXPAR) that combines the benefits of Cas9/sgRNA site-specific cleavage and the rapid amplification kinetics of EXPAR (Figure 2) [34]. Sensitive DNA detection with CAS-EXPAR could be achieved in 1 h by combining a real-time fluorescence intensity analysis approach. A dCas9-based amplification process, developed by Wang et al., is a simple but innovative nucleic acid amplification approach (Cas9nAR) [19]. The intercalation dye SYBR Green I was used to track Cas9nAR progression in real-time, and the fluorescence intensity is proportional to the concentration of dsDNA products.
It is found that RNA-guided DNA binding triggers Cas12a’s indiscriminate trans-cleavage activity [12]. Cas13a, such as Cas12a, possesses the trans-cleavage ability for non-specific nucleic acids after binding to target RNA [11]. Yet, Cas13a completely cuts non-target RNA following the creation of the guide-target RNA duplex. Similar to Cas12a, Cas14 detects a target and initiates non-specific ssDNA trans-cleavage activity [10]. Based on this principle, fluorescence resonance energy transfer (FRET) reporters with one fluorophore acting as an energy source or fluorophore and the other as a receptor at both ends of ssDNA or RNA are designed in CRISPR biosensors [59,60,61]. It begins to cleave these FRET reporters non-specifically when Cas enzymes start trans-cleavage activity due to the recognition of targets. As we all know, with one fluorophore acting as an energy source or fluorophore and the other as a receptor, FRET is a distance-dependent mechanism of energy transfer between fluorophores and quenchers [62]. The length of the fluorophore/quencher labeled ssDNA reporter is an important factor affecting analytical performance [63,64]. When ssDNA reporters are cleaved by Cas enzymes, the reporter dye is separated from the quencher dye. The quenching effect is gone, and thus reporter dye fluorescence will be detected by the instrument. The released reporter dye signal is proportional to the concentration of targets. The DNA endonuclease-targeted CRISPR trans-reporter (DETECTR) approach provides attomolar sensitivity for DNA detection by combining Cas12a ssDNase activation with isothermal amplification. Li et al. employed PCR or other isothermal amplification techniques to specifically amplify the target DNA [31]. The amplicon was then combined with the Cas12a/crRNA complex, and a ternary complex was formed when the target DNA was present. The quenched fluorescent ssDNA reporter was trans-cleaved upon the formation of the ternary complex, releasing the fluorescence. In contrast to Cas12a, which had a greater nonspecific ssDNA trans-cleavage rate with target dsDNA than with target ssDNA, Cas12b demonstrated a distinct target preference in trans-cleavage. By using the Cas12b, a thermophilic RNA-guided endonuclease from the type V-B CRISPR/Cas system, Li et al., created HOLMESv2 for discriminating single nucleotide polymorphism (SNP) [48]. It was also combined the nucleic acid amplification and target identification phases into a single system to streamline operations and prevent cross-contamination. The needed target concentration was reduced to 10−8 nM when paired with loop-mediated isothermal amplification (LAMP), which was equivalent to Cas12a. The FRET reporters used in Cas12a studies include 5(6)-carboxyfluorescein (FAM) as the donor and BHQ1 as the quencher. Although the ssDNA reporters are commonly used for research and applications (Table 2), limiting selections of ssDNA reporters from an application and probe diversification standpoint. We recently investigated the properties of various types of fluorescent probes for CRISPR in detail (Figure 3) [65]. We demonstrated that the trans-cleavage of Cas12a is not limited to ssDNA or dsDNA reporters but can be extended to molecular beacons (MB). MBs are a class of small, single-stranded oligonucleotides with a fluorophore and a quencher at the two termini, which have been widely used as hybridization-activated FRET probes but rarely used in CRISPR-based fluorescent assays. On the 5′ and 3′ termini of 37-nt ssDNA, we labeled the MB reporter with Taxes Red and BHQ2. A secondary structure such as the hairpin shape (15-nt in the loop structure and 12-bp in the stem structure) is formed after annealing. Results demonstrated that MB probes can achieve better analytical performance than ssDNA and dsDNA probes and that FRET probes modified with a fluorophore (Texas Red) are more sensitive than modified FAM. Accordingly, we developed a highly sensitive SARS-CoV-2 detection with a sensitivity as low as 100 fM, which was successfully achieved without an amplification strategy. A real coronavirus, GX/P2V instead of SARS-CoV-2, was chosen for practical application validation. After magnetic bead-based rapid RNA extraction and RT-PCR amplification, a minimum of 2.7 × 101 copies/mL can still be obtained. The inspiration can also apply to other Cas effectors with trans-cleavage activity, which provides perspectives for simple, highly sensitive, and universal molecular diagnosis in various applications. Numerous efforts have been devoted to RNA virus detection based on the trans-cleavage activity of RNA reporters of Cas13a after target RNA recognition (Table 3). For example, Gootenberg et al. designed a Cas13a-mediated in vitro nucleic acid detection platform named Specific High-Sensitivity Enzymatic Reporter UnLOCKing (SHERLOCK), which uses commercially available RNA fluorescent probes as fluorescent signals [21]. The dsDNA samples or RNA samples were amplified by recombinase polymerase amplification (RPA) or RT-RPA, and then T7 RNA polymerase transcribes the amplified DNA to RNA. The target RNA products trigger the trans-cleavage activity of Cas13 and cause the fluorescent reporters to be cleaved to release fluorescent signals. Sherlock Biosciences, Inc. (Watertown, MA, USA) has received EUA (Emergency Use Authorization) approval in the United States for a fluorescent CRISPR kit. RNA is extracted from clinical samples using the PureLink™ Viral RNA/DNA Mini Kit. The viral RNA is then reverse transcribed and amplified by RT-LAMP, which activates the cleavage activity of the CRISPR complex to cut reporters, thereby releasing a fluorescent light signal that can be observed by the plate reader. To make the actual operation easier, Myhrvold et al., developed HUDSON, a method to lyse viral particles and deactivate the large amounts of ribonucleases found in body fluids using heat and chemical reduction, to identify viral nucleic acid directly from bodily fluids via SHERLOCK (Figure 4) [21,22]. Without dilution or purification, HUDSON-treated urine or saliva can be introduced directly into the RPA reaction mixture. Despite lower viral titers than those in serum, HUDSON and SHERLOCK enabled sensitive and specific ZIKV nucleic acid detection in 1 h. The Cas14 system is the smallest functional CRISPR system discovered to date, about one-third the size of Cas9, and features PAM sequence independence and high specificity and fidelity during shearing. Harrington et al. used Cas14 to design the DETECTR platform by combining the isothermal amplification method of RPA with the high-fidelity detection of DNA by Cas14 [10]. The amplified target DNA induced Cas14 to cleave the fluorescent reporter ssDNA, resulting in fluorescent signal recovery. When Cas14a was incubated with various sgRNA or target ssDNA, it showed a strong preference for longer ssDNA fluorescent substrates. The 12 T-base ssDNA reporter was the first to show the strongest fluorescent signal [70]. Wei et al. were the first to discover that Cas14a has trans-cleavage of ssDNA, specifically activated by a target RNA, without self-destruction [28]. Accordingly, they created a series of target RNA and ssDNA with varying lengths and mutation sites to investigate the effect of the active factor (target RNA or ssDNA) on Cas14a’s trans-cleavage activities. It was discovered that the Cas14a/sgRNA-target RNA self-assembled complex can only trans-cleave the ssDNA reporter (Table 4). Furthermore, when triggered by target RNA, the substrate selectivity of this complex can be higher than when activated by target ssDNA. Based on this finding, a new ATCas-RNA platform was created that can detect pathogens using a FRET reporter with excellent selectivity and sensitivity, similar to the Cas13-based RNA diagnostic system.
It is a continuous challenge to improve the detection limit and sensitivity of assays for capturing and properly estimating the concentration of a certain target. Traditional fluorescent methods for concentration measurement are based on a bulk solution response. The digital biosensor is a game-changing technique for both analytical chemistry and single-molecule studies. The most notable distinction between digital bioassays and other analytical methods is that the former counts just the number of a binary (positive or negative) signal from a vast population of individual reactors for concentration determination, whereas the latter records the absolute signal strength from a single reactor. Digital bioassays work by separating reaction solutions into micrometer-sized compartments [71]. In digital bioassays, rather than quantifying the absolute intensity of ensemble signals from tubes or microtiter plates, just the fraction of microcompartments displaying positive signals is tallied. The bulk solution enzyme concentration is calculated by dividing the enzyme-containing solution and a suitable substrate into a large number of femtoliter-sized reactor containers. The reactor volumes are small enough that they can only hold zero or one enzyme molecule. A binary readout approach can be used to count enzyme molecules by observing the presence or absence of a fluorescent product arising from single enzyme molecule catalysis in each reaction vessel. Digital biosensors combined with excellent CRISPR enzyme properties would create new sparks. Park et al. designed and applied the first digital CRISPR/Cas-assisted assay—digitization-enhanced CRISPR/Cas-assisted one-pot virus detection (deCOViD)—to SARS-CoV-2 detection (Figure 5) [44]. The DeCOViD was achieved by adjusting and discretizing a one-step CRISPR/Cas12a-assisted RT-RPA into 0.7 nL digital reaction wells within commercially available microfluidic digital chips. With a high signal-to-background ratio, broad dynamic range, and high sensitivity, deCOViD can achieve qualitative detection in 15 min and quantitative detection in 30 min using evenly elevated digital concentrations. SARS-CoV-2 RNA and inactivated SARS-CoV-2 can be fluorescently identified in deCOViD using a single-step technique that combines RPA and CRSIPR/Cas12a-based detection. RNA targets are reverse transcribed and amplified into DNA amplicons via RT-RPA, which activates Cas12a guide RNA complexes, which cleave ssDNA fluorogenic reporters and produce fluorescence. This single-step assay is then fine-tuned to guarantee that it can be loaded quickly onto commercially available microfluidic digital chips and discretized reliably within digital reaction wells. During assay digitization, each copy of the target is separated at a locally elevated concentration inside digital reaction wells, allowing for rapid amplification regardless of sample concentration. This beneficial method improved the signal-to-background ratio, dynamic range, and sensitivity. One-pot digital warm-start CRISPR (WS-CRISPR) reaction was partitioned into sub-nanoliter aliquots using QuantStudio 3D digital chips to create the digital WS-CRISPR assay for detecting SARS-CoV-2 in clinical COVID-19 samples with high sensitivity by Ding and coworkers [47]. The WS-CRISPR reaction combined low-temperature reverse transcription dual-priming isothermal amplification (RT-DAMP) and CRISPR/Cas12a-based fluorescence detection in a one-pot format and is efficiently initiated at temperatures above 50 °C, preventing premature target amplifications at room temperature and allowing accurate nucleic acid digital quantification. In one tube, a one-pot WS-CRISPR reaction mixture was first prepared. Over ten thousand sub-nanoliter (0.7 nL) microreactions are isolated in microwells after being dispersed into the QuantStudio 3D digital chip. Each microreaction with SARS-CoV-2 RNA target undergoes WS-CRISPR reaction and generates intense green fluorescence (positive spots) when incubated at 52 °C, whereas those without target do not (negative spots). The digital WS-CRISPR assay was designed to quantitatively identify 32 clinical swab samples and three clinical saliva samples by targeting the SARS-CoV-2 nucleoprotein (N) gene. Furthermore, the digital WS-CRISPR test can detect SARS-CoV-2 in heat-treated saliva samples without the need for RNA extraction. Digital bioassays are compatible with existing classical techniques widely used in normal biological, chemical, or clinical laboratories. Ning et al. described a COVID-19 CRISPR-FDS assay that does not require any specific equipment beyond what is commonly available in research and clinical settings [45]. RNA extraction, target amplification, and fluorescence signal detection are the three processes of CRISPR-FDS. On a 96-well half-area plate, the gRNA/Cas12a complex, which is regulated by a target-specific synthetic gRNA, recognized the target amplicon, and caused it to non-specifically cleave a reporter oligo modified with fluorescein and a quencher molecule at each terminal, resulting in the release of fluorescence. Combined CRISPR/Cas13-based RNA recognition with microchamber-array technologies, Shinoda et al., developed a CRISPR-based amplification-free digital RNA detection named SATORI [51]. Amplification-free RNA molecular detection at the single-molecule level was achieved by using a device containing more than 1,000,000 through-holes in a femtoliter microchamber. The microchamber device is a glass block with an intake port mounted to a glass substrate with 1,000,000 microchambers, with a U-shaped spacer seal in between. The pre-assembled Cas13-crRNA complexes were mixed with FRET reporters, target RNA, and loaded into the microchamber device that were used in SHERLOCK. After the sealing of the device, a fluorescent signal released from the cleaved FRET reporters can be observed by fluorescence microscopy. Furthermore, Ackerman et al. developed a highly multiplexed nucleic acid detection platform called Carmen (Figure 6) [53]. Each nucleic acid sample is amplified by RPA or PCR and combined with a unique, solution-based fluorescent color code that acts as an optical identifier. In the fluorous oil, each color-coded solution was emulsified to produce 1 nL droplets. The Cas13-detection mix, which contained commercially available quenched fluorescent RNA reporter, a sequence-specific crRNA, and Cas13, was operated in the same way as the nucleic acid sample. The microwell-array chip is made of polydimethylsiloxane (PDMS), which has a large number of microwells that each of them can hold only two droplets. A variety of different droplets can be mixed, and after two droplets are randomly filled in the microwells of the chip, the contents of the two droplets added to the microwells can be determined by observing and identifying the color code of the droplets through a fluorescence microscope. If the nucleic acid sample in the droplet triggers the trans cleavage activity of Cas13, it will cause the RNA reporter to be cut and release fluorescence, and the corresponding detection result can be derived from the color of the mixed-droplet.
Nanomaterials have distinct physical and chemical characteristics, making them ideal candidates for the development of innovative fluorescent biosensors. The enormous potential of such fluorescent nanomaterials has opened the way for the development of novel biomolecule assays with high analytical capabilities, such as sensitivity, cost-effectiveness, and simplicity of use. Various fluorescent nanomaterials with excellent fluorescence characteristics, such as QDs, photonic crystals (PHC), and metal nanoparticles, have been widely employed for CRISPR biosensors. In this section, we will focus on the detection methods of different Cas enzymes using fluorescent nanomaterials as signal labels through CRISPR fluorescent measurement.
QDs, as fluorescent semiconductor nanocrystals, have excellent photoluminescent quantum yields and great photochemical stability. They feature broad absorption and a narrow and symmetric photoluminescence spectrum (from UV to near-infrared) that may be modified by altering the size and chemical composition of the nanocrystals. The excellent optical properties make QDs the ideal candidates for CRISPR biosensors. In this context, Zhou et al., developed a CRA-LFB assay using a new CRISPR/Cas12a-based fluorescence enhanced lateral flow biosensor (LFB) in combination with functionalized QDs and recombinase-assisted amplification (RAA) [72]. The Cas12a-mediated trans-cleavage activation generated by the target DNA to digest biotin-DNA probes was seen in the CRA-LFB assay, which had no complementarity with the capture probe mounted on the test (T) line and resulting in an undetected T line fluorescence signal on LFB. The fluorescence intensity of the T and control lines could be measured using the naked eye or a fluorescence strip reader. This assay can detect S. aureus in both spiked and natural meat and vegetable samples.
Photonic crystal barcodes (PhC) are artificial periodic dielectric structures with photonic bandgap properties and unique optical phenomena such as suppressed spontaneous emission, highly reflective omnidirectional mirrors, and low-loss waveguides. Due to its exceptional encoding stability and resistance to interference from fluorescent background or photo-bleaching, PhC barcodes, which may be created via the self-assembly of monodisperse colloidal nanoparticles, have caught the attention of scientists. These properties make PhC barcodes a promising platform for CRISPR biosensors without disrupting the fluorescent signal. For example, Zhang et al. developed a CRISPR/Cas9 technology for multiplex and sensitive nucleic acid detection based on bioinspired PhC barcodes [41]. Because of the polydopamine (PDA) coating, the bioinspired PhC barcodes feature not only discrete structural color as encoding components but also rich functional surface groups for probe immobilization. The CRISPR/Cas9 system detected and cleaved target DNA, generating ssDNA with the assistance of the Klenow fragment, which was subsequently captured by PhC barcodes, and the detecting signal was amplified by a hybridization chain reaction (HCR). Different colors of colloidal crystal barcodes were employed for multiplexed detection, with the fluorescence signal only appearing on the colloidal crystal particles capturing the matching color. It was able to detect Human Papilloma Virus (HPV) nucleic acids with a sensitivity limit of 0.025 pM and be multiplexed assayed with high accuracy and specificity, providing a fresh insight for multiplexed biomarker quantification, and showing great potential in clinical disease diagnostics.
Nanometallic materials are metals and alloys that form nanocrystalline grains. It has the characteristics of a grain boundary ratio, specific surface energy, and a large ratio of surface atoms. Noble metal nanomaterials (NMN) benefit from localized surface plasmon resonance (LSPR), which can be used as a highly sensitive signal conversion unit, thus enabling highly sensitive assay. When the distance between the fluorophore and the NMNs is less than about 2 nm, NMNs can function as fluorescence quenchers. This property can be used to improve the primary signal intensity when detecting biomaterials with NMNs such as Au and Ag nanomaterials. Utilizing DNA-functionalized Au nanoparticles (AuNPs) and metal-enhanced fluorescence (MEF) effects, Choi et al. developed a cfDNA detection biosensor based on CRISPR/Cas12a nucleic acid amplification (Figure 7) [30]. MEF happened when the target cfDNA activated the Cas12a complex, followed by ssDNA degradation between AuNP and fluorophore, resulting in color shifts from purple to red-purple. Breast cancer gene-1 was detected with great sensitivity in 30 min with this technique. This rapid and highly selective sensor could be utilized in POCT scenarios to assess nucleic acid indicators such as viral DNA.
Since fluorescence needs an excitation light source, portable signal reading is not easy to achieve in general. Recently, many researchers have developed small fluorescence spectrometers for CRISPR biosensors by using a laser or LED as a light source and the naked eye or smartphone as a detector, which is convenient for point-of-care (POC) tests [73,74,75]. Naked-eye detection can get rid of the use of large instruments and the involvement of professionals and does not even require specialized skills training. The test can be performed under very simple conditions and the results can be known by the change of color. Fluorescence-based CRISPR/Cas systems still require instruments to read out the signal and do not allow true naked-eye testing. However, by exploiting the special optical properties of gold nanoparticles and endowing them with certain functions, naked-eye detection of CRISPR/Cas systems can be achieved [30]. In addition, smartphones, which are almost a must-have, are also a convenient option for signal readout. The combination of smartphones and cleverly designed microdevices with CRISPR/Cas technology greatly enhances the convenience and safety of detection, reducing the difficulty of the operation and the risk of centralized cross-infection [76]. In the statistics, summary, query, and other aspects of data also have a huge potential for development and broad application prospects. Using a fluorescence-based POC system to combine potent CRISPR/Cas assay with a fluorescence-based POC system for speedy and accurate virus detection, He et al., reported a method for African swine fever virus (ASFV) target DNA detection [77]. The POC system has a fluorescence sensing device with a 488 nm laser as an excitation source by self-designed compact detection equipment. The small fluorescence-sensing unit is aligned with an 80× disposable cartridge, thus matching the high-throughput of commercialized PCR systems. The Cas12a/crRNA/ASFV DNA complex is activated when ASFV DNA is bound, and it destroys a fluorescent ssDNA reporter contained in the test. The fluorophore on the ssDNA probe was released into the assay and detected by the fluorescence-sensing unit. Furthermore, using a blue LED or UV light illuminator, Ding et al. designed an all-in-one dual Cas12a (AIOD-CRISPR) test for SARS-CoV-2 detection that is simple, quick, ultrasensitive, selective, one-pot, and visible [43]. The Cas12a endonuclease was activated when the Cas12a-crRNA complexes bind the target sites, cleaving the FRET reporters and generating strong fluorescence signals. After incubation, the reaction mixture generated a super-bright fluorescence signal that could be observed clearly with a blue LED or UV light illuminator. The color of the reaction tube transition from orange-yellow to green could be seen with the naked eye even in the absence of stimulation under ambient light circumstances. All components for nucleic acid amplification and CRISPR-based detection are thoroughly mixed in a single, one-pot reaction system, obviating the requirement for separate pre-amplification and amplified product transfer. Combined an automated and multiplexing CRISPR microfluidic chip with a custom-designed benchtop fluorometer for rapid- and low-volume virus detection, Qin et al., used Cas13a’s collateral RNA destruction after activation to establish an automated POC method for Ebola RNA detection [78]. Following automated microfluidic mixing and hybridization, nonspecific cleavage products of Cas13a are quickly analyzed by a patented integrated fluorometer, which is small in size and suitable for in-field diagnostics. The microfluidic chip is mounted on the fluorometer for in situ detection. By using a mobile phone camera to measure fluorescence within a compact device that includes low-cost laser illumination and collection optics. Fozouni et al. developed an amplification-free CRISPR/Cas13a assay for direct detection of SARS-CoV-2 from nasal swab RNA that can be read with a mobile phone microscope [76]. By combining multiple crRNAs to increase Cas13a activation then cleave any RNAs in the vicinity indiscriminately, detected using a fluorophore-quencher pair coupled by an RNA that fluoresces when active Cas13 cleaves it. The high sensitivity of mobile phone cameras, together with their connectivity, GPS, and data-processing capabilities, have made them attractive tools for point-of-care disease diagnosis in low-resource regions.
A comprehensive understanding of various fluorescence readout types and variables affecting their analytical performance can facilitate superior experimental designs. This mini-review summarized different fluorescence readout types of the CRISPR/Cas system, as well as recent representative progress and applications. The fluorescence intensity was measured using dsDNA-specific fluorophores of cas9 and dcas9 based on binding activity. The fluorescence recovery was measured using FRET reporters for cas12, cas13, and cas14 based on trans cleavage activity. With the advancement of CRISPR technology, the CRSIPR/Cas system, with its distinct characteristics, will provide infinite possibilities in the future. Despite significant advancements and appealing features currently, many challenges remain. It is currently difficult to achieve multiplexed detection in a single sample based on trans-cleavage activity by measuring the overall fluorescence intensity signal. According to a fluorescent digital principle similar to digital PCR, some researchers exploited the chip with adequate nanopores to achieve high-sensitivity digital CRISPR detection. Multiplexed detection is realized easily using space separation and fluorescent color coding. Furthermore, naked-eye detection can be achieved by test strips or simplified light sources, such as laser, LED, or smartphone light, which is appropriate for POCT settings with limited resources. However, CRISPR biosensors are dependent on signal amplification such as PCR, RPA, and LAMP presently. How to achieve amplification-free CRISPR detection using signal readout techniques is a major current challenge. The advancement of nanotechnology opens up new avenues to enrich the CRISPR signal readout library. More signal readout techniques combined with nanomaterials to achieve simpler and more sensitive analytical performance are a future trend for diverse applications of unique features of the CRISPR/Cas system. In addition, how to design the signal readout mode to realize multiplex detection is also a major challenge. Although challenges still remain, CRISPR/Cas-based nucleic acid detection will have broad applications, especially in the SARS-CoV-2 and other pathogenic diseases or cancers diagnosis. | true | true | true |
|
PMC9599952 | Elena Andreucci,Jessica Ruzzolini,Francesca Bianchini,Giampaolo Versienti,Alessio Biagioni,Matteo Lulli,Daniele Guasti,Patrizia Nardini,Simona Serratì,Francesca Margheri,Anna Laurenzana,Chiara Nediani,Silvia Peppicelli,Lido Calorini | miR-214-Enriched Extracellular Vesicles Released by Acid-Adapted Melanoma Cells Promote Inflammatory Macrophage-Dependent Tumor Trans-Endothelial Migration | 18-10-2022 | acidic tumor microenvironment,melanoma,extracellular vesicles,miR-214,inflammation,vascular permeability,trans-endothelial migration | Simple Summary Cutaneous melanoma is the most aggressive form of skin cancer with high-metastatic ability. Despite the recent advancements in melanoma treatments, the prognosis of metastatic patients remains very poor. A better understanding of the molecular mechanisms leading to melanoma dissemination is urgently needed in order to develop novel therapeutical strategies to ameliorate patients’ outcomes. Extracellular vesicles (EV) released by tumor cells are key players in metastasis development: by conveying bioactive molecules with oncogenic activity, they can modulate the surrounding—and even the distant—microenvironment and reprogram recipient cells to facilitate the metastatic cascade. Here, we show that melanoma cells release a higher amount of miR-214-enriched EV when adapted to extracellular acidosis, which promote a macrophage activation state, capable of facilitating the trans-endothelial migration of melanoma cells. Thus, we disclose a new molecular mechanism to prevent melanoma intravasation based on miR-214 targeting. Abstract The understanding of the molecular mechanisms leading to melanoma dissemination is urgently needed in view of the identification of new targets and the development of innovative strategies to improve patients’ outcomes. Within the complexity of tumor intercellular communications leading to metastatic dissemination, extracellular vesicles (EV) released by tumor cells are central players. Indeed, the ability to travel through the circulatory system conveying oncogenic bioactive molecules even at distant sites makes EV capable of modulating recipient cells to facilitate metastatic dissemination. The dynamic remodeling of the tumor microenvironment might influence, along with a number of other events, tumoral EV release. We observed that, in melanoma, extracellular acidosis increases the release of EV enriched in miR-214, an onco-miRNA involved in melanoma metastasis. Then, miR-214-enriched EV were found to induce a state of macrophage activation, leading to an overproduction of proinflammatory cytokines and nitric oxide. Such an inflammatory microenvironment was able to alter the endothelial cell permeability, thereby facilitating the trans-endothelial migration of melanoma cells, a crucial step in the metastatic cascade. The use of synthetic miR-214 inhibitors and miR-214 overexpression allowed us to demonstrate the key role of miR-214 in the EV-dependent induction of macrophage activation. Overall, our in vitro study reveals that the release of tumor miR-214-enriched EV, potentiated by adapting tumor cells to extracellular acidosis, drives a macrophage-dependent trans-endothelial migration of melanoma cells. This finding points to miR-214 as a potential new therapeutic target to prevent melanoma intravasation. | miR-214-Enriched Extracellular Vesicles Released by Acid-Adapted Melanoma Cells Promote Inflammatory Macrophage-Dependent Tumor Trans-Endothelial Migration
Cutaneous melanoma is the most aggressive form of skin cancer with high-metastatic ability. Despite the recent advancements in melanoma treatments, the prognosis of metastatic patients remains very poor. A better understanding of the molecular mechanisms leading to melanoma dissemination is urgently needed in order to develop novel therapeutical strategies to ameliorate patients’ outcomes. Extracellular vesicles (EV) released by tumor cells are key players in metastasis development: by conveying bioactive molecules with oncogenic activity, they can modulate the surrounding—and even the distant—microenvironment and reprogram recipient cells to facilitate the metastatic cascade. Here, we show that melanoma cells release a higher amount of miR-214-enriched EV when adapted to extracellular acidosis, which promote a macrophage activation state, capable of facilitating the trans-endothelial migration of melanoma cells. Thus, we disclose a new molecular mechanism to prevent melanoma intravasation based on miR-214 targeting.
The understanding of the molecular mechanisms leading to melanoma dissemination is urgently needed in view of the identification of new targets and the development of innovative strategies to improve patients’ outcomes. Within the complexity of tumor intercellular communications leading to metastatic dissemination, extracellular vesicles (EV) released by tumor cells are central players. Indeed, the ability to travel through the circulatory system conveying oncogenic bioactive molecules even at distant sites makes EV capable of modulating recipient cells to facilitate metastatic dissemination. The dynamic remodeling of the tumor microenvironment might influence, along with a number of other events, tumoral EV release. We observed that, in melanoma, extracellular acidosis increases the release of EV enriched in miR-214, an onco-miRNA involved in melanoma metastasis. Then, miR-214-enriched EV were found to induce a state of macrophage activation, leading to an overproduction of proinflammatory cytokines and nitric oxide. Such an inflammatory microenvironment was able to alter the endothelial cell permeability, thereby facilitating the trans-endothelial migration of melanoma cells, a crucial step in the metastatic cascade. The use of synthetic miR-214 inhibitors and miR-214 overexpression allowed us to demonstrate the key role of miR-214 in the EV-dependent induction of macrophage activation. Overall, our in vitro study reveals that the release of tumor miR-214-enriched EV, potentiated by adapting tumor cells to extracellular acidosis, drives a macrophage-dependent trans-endothelial migration of melanoma cells. This finding points to miR-214 as a potential new therapeutic target to prevent melanoma intravasation.
Cutaneous melanoma is the most lethal form of skin cancer, accounting for more than 320,000 cases and about 57,000 cancer-related deaths per year worldwide [1]. The incidence of melanoma is globally increasing more rapidly than any other kind of cancer [2]. If not removed early by surgical resection, cutaneous melanoma disseminates very rapidly leading to metastatic disease and poor patient outcomes. The 5-year survival rate for patients affected by stage IV melanoma is dramatically nearby 15%, and despite the recent introduction of new therapeutic options, such as immuno- and targeted therapy, which has significantly decreased the mortality rate, much is still needed to make melanoma curable [3]. Extracellular vesicles (EV) are crucial mediators within the complex array of cell–cell interactions involved in the several steps of the metastatic cascade. EV are nano-sized membranous structures released by most cells into the extracellular space, including body fluids [4]. Based on their size, EV are classified as exosomes (diameter 30–100 nm), microvesicles (also called ectosomes; 100–1000 nm), and apoptotic bodies (1000–5000 nm) [5]. EV are involved in several biological functions, such as the removal of harmful cellular materials, trophic support, and cell–cell communication. Indeed, the ability of EV to travel through body fluids—conveying functional information to distant sites in vivo—and infiltrate the biological barriers make them capable to drive intercellular communication in all tissues in both physiologic and pathologic conditions [6]. This not only changed completely the concept of intercellular communication but also clarified several cellular processes in cancer [7]. EV cargos comprise classical soluble and insoluble signaling factors, such as structural proteins, lipids, and nucleic acids, including microRNAs (miRNAs). miRNAs are small single-stranded RNAs, with an average of 22 nucleotides in length, belonging to the family of non-coding RNAs. miRNAs affect DNA, RNA, and proteins acting as negative regulators of gene expression, being able to control the translation and even the transcription of target mRNAs. Actually, miRNAs act as molecular sponges, influencing multiple biological processes, including cell proliferation, differentiation, migration, angiogenesis, and apoptosis, and their dysregulated expression is associated with a plethora of diseases, including cancer [8]. All of this makes miRNAs an attractive target for oncology research [9]. Among the melano-miRNAs (i.e., miRNAs involved in melanoma formation and progression), miR-214 has been found to promote melanoma metastasis by increasing migration, invasion, extravasation, and survival of melanoma cells into the circulatory system [10,11,12]. Notably, miR-214 targeting has been recently proposed as a promising anti-metastatic therapy [13]. Here, we show that miR-214, conveyed by EV particularly released when melanoma cells were adapted to extracellular acidosis, exerts a proinflammatory activity able to alter the vascular structures to facilitate the trans-endothelial migration of melanoma cells. Extracellular acidosis is now recognized as a crucial aspect of the tumor microenvironment (TME), which is central in cancer progression, including melanoma. These findings provided mechanistic insight into the role of miR-214 in melanoma progression, strengthening the possibility to exploit miR-214 targeting to prevent or treat metastatic disease.
B16-F10 (hereafter referred to as B16) were kindly supplied by Dr S. Gattoni-Celli (Medical University of South Carolina, Charleston, SC, USA) [14]. B16-LU and B16-LI murine melanoma cell lines were previously isolated in our laboratory from B16-derived lung and liver metastasis, respectively, in a syngeneic metastatic experimental model. Briefly, B16-LU and B16-LI were obtained by the so-called experimental metastasis assay consisting of the intravenous injection of B16 murine melanoma cells into C57Bl/6 syngeneic mice. The primary A375 and Sk-Mel-2 cell lines, the metastatic WM-266-4 human melanoma cell line, and the murine macrophage RAW 264.7 cell line were purchased by American Type Culture Collection (ATCC; Rockville, MD, USA). The murine endothelial cells were kindly given by Dr. Ilaria Cimmino and Professor Francesco Oriente (Department of Translational Medicine, Research Unit (URT) Genomic of Diabetes, Institute of Experimental Endocrinology and Oncology, National Council of Research (CNR), University of Naples Federico II, Naples, Italy). All the cell lines used were grown at a 37 °C, 5% CO2 humidified atmosphere in DMEM 4.5 g/L glucose, 2 mM L-glutamine, and 10% FBS (Euroclone, Pero, Italy) (decomplemented FBS was used for culturing RAW 264.7 macrophages). Extracellular acidosis was mimicked in vitro by culturing melanoma cells in pH 6.7 ± 0.1 medium for at least 3 months before use, as previously described [15]. MISSION® Synthetic microRNA Inhibitors anti-miR-214-3p and anti-miR-214-5p were purchased by Merck Life Science S.r.l. and administered to RAW 264.7 macrophages in the presence of EV to interfere with miR-214 activity. miR-214 Mouse MicroRNA Expression Plasmid and the empty vector were purchased by OriGene Technologies GmbH (Herford, Germany) and transfected through Lipofectamine™ 3000 Transfection Reagent (Thermo Fisher Scientific, Milan, Italy) in RAW 264.7 cells. The selection of miR-214-overexpressing cells (miR-214+) was performed by geneticin treatment (Merk Life Science S.r.l., Milan, Italy) followed by cell sorting for GFP with a BD FACS Melody (BD Biosciences, Milan, Italy) flow cytometer. Lipopolysaccharide (LPS, 10 ng/mL) was administered to RAW 264.7 macrophages as a positive control of the proinflammatory macrophage activation. 1 × 106 RAW 264.7 cells were seeded into 6-well plates; treated or not for 16 h with EV ± anti-miR-214 or LPS; and conditioned media (CM) was collected for either IL-1β, IL-6, and TNF-α ELISA or nitric oxide (NO) detection (in this case, DMEM 4.5 g/L glucose without phenol red was used) or for a 24-h treatment of murine endothelial cells.
Melanoma cells were starved from FBS for 24 h, and 30 mL of the culture medium was collected and subjected firstly to 300× g for 5 min to discard the cell pellet and then to 3000× g for 30 min at 4 °C to remove cell debris and apoptotic bodies. The supernatant was then subjected to 10,000× g at 4 °C for 1 h and 100,000× g at 4 °C for 1 h to isolate the ectosome and exosome fractions, respectively. Based on their final destination, the ectosome and exosome fractions were resuspended in PBS or culture medium and mixed in a single EV-containing solution.
Samples were prepared for transmission electron microscopy (TEM) by the negative and positive staining procedures. In brief, the EV mixtures comprising exosomes and ectosomes were fixed in Karnovsky’s fluid for 5 min at room temperature (RT), centrifuged for 5 min at 11,000× g, and then rinsed and resuspended in Cacodylate buffer 0.2 M. Aliquots of these suspensions were sedimented for 5 min at RT or 15 min at 37 °C on 300 mesh, and copper/formvar-coated grids depending on the kind of staining (respectively, negative and positive). The Uranyless (Electron Microscopy Sciences, Hatfield, PA, USA) was used both as negative and positive staining for 5 RT and 15 min at 37 °C. Dried samples were analyzed at a Jem 1010 transmission electron microscope (Jeol, Tokyo, Japan) at 80 Kv and photomicrographs acquired with a MegaView III digital camera (Soft Imaging System, Muenster, Germany) and analyzed with AnalySIS software (Soft Imaging System, Muenster, Germany).
A nanoparticle tracking analysis (NTA) was performed, as previously described [16], with a NanoSight NS300 (Malvern Panalytical, Westborough, MA, USA) apparatus equipped with a 488-nm excitation laser and an automated syringe sampler. According to the NanoSight technology, the EV size is calculated through the Stokes–Einstein equation based on the relationship between the Brownian motion and hydrodynamic diameter. The EV samples were diluted at 1:500 in PBS and loaded into 1-mL syringes. CSV files generated by NTA by software v3.2 were used for a computational analysis.
Total RNA was prepared using Tri Reagent (Merk Life Sciences S.r.l., Milan, Italy), agarose gel checked for integrity, and quantified at the NanoDrop™ 8000 Spectrophotometer (Thermo Fisher Scientific, Milan, Italy). Reverse transcription was performed with the MystiCq® microRNA cDNA Synthesis Mix (Merck Life Science S.r.l., Milan, Italy), and miR-214 expression was evaluated at the Bio-Rad CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad, Milan, Italy) using MystiCq® microRNA® SYBR® Green qPCR ReadyMix™ (Merck Life Science S.r.l., Milan, Italy) according to the manufacturer’s instructions. MystiCq® miR-214-3p/5p and qPCR Assay Primers and MystiCq® microRNA qPCR Control Primer (RNU-6) were purchased from Merck Life Science S.r.l.
Total RNA was prepared using Tri Reagent (Merk Life Sciences S.r.l., Milan, Italy), agarose gel checked for integrity, and quantified with the NanoDrop™ 8000 Spectrophotometer (Thermo Fisher Scientific, Milan, Italy). Reverse transcription was performed with the iScript cDNA Synthesis Kit (Bio-Rad, Milan, Italy) according to the manufacturer’s instructions. Selected genes were evaluated at the Bio-Rad CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad, Milan, Italy) using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, Milan, Italy). The primer sequences (Merck Life Science S.r.l., Milan, Italy) are the following: IL-1β FW: CCT GCA GCT GGA GAG TGT GGA; IL-1β RV: CCC ATC AGA GGC AAG GAG GAA; IL-6 FW: CTT CCA TCC AGT TGC CTT CT; IL-6 RV: TGC ATC ATC GTT GTT CAT AC; TNF-α FW: GCG GTG CCT ATG TCT CAG CC; TNF-α RV: TGA GGA GCA CGT AGT CGG GG; 18S FW: CGC CGC TAG AGG TGA AAT TCT; 18S RV: CGA ACC TCC GAC TTT CGT TCT; β-ACTIN FW: CAT TGC TGA CAG GAT GCA GAA GG; β-ACTIN RV: TGC TGG AAG GTG GAC AGT GAG G.
RAW 264.7 cells were seeded on glass coverslips in six-well plates and treated with EV isolated from B16, B16-LU, and B16-LI cell lines under control pH or chronic acidic conditions. RAW 264.7 macrophages were fixed for 30 min at 4 °C with 3.7% paraformaldehyde and permeabilized for 15 min with PBS 0.1% Triton X-100 at RT. After 1-h incubation in blocking buffer (0.1% Triton X-100 and 5.5% horse serum PBS), cells were stained with anti-NF-kB p65 (GeneTex, purchased from Prodotti Gianni, Milan, Italy) for 1 h at RT and then 45 min at RT in the dark with Cy3-conjugated anti-mouse antibody (Thermo Fisher Scientific, Milan, Italy). Following the 20-min nuclei staining with DAPI (Thermo Fisher Scientific, Milan, Italy) at RT in the dark, cells were mounted onto glass slides and visualized at the SP8 confocal microscope (Leica Microsystems, Milan, Italy). Manders coefficients by ImageJ software, proportional to the amount of fluorescence of the colocalizing pixels in NF-ĸB and DAPI color channels, were used to calculate the NF-ĸB nuclear fraction.
Melanoma-derived EV and RAW 264.7 macrophages were lysed in radioimmunoprecipitation assay (RIPA) lysis buffer (Merck Millipore, Milan, Italy) added with Pierce Protease Inhibitor Tablets (Thermo Fisher Scientific, Milan, Italy) for protein isolation. The protein concentration was measured with Bradford reagent (Merck Millipore, Milan, Italy), and equal amounts of protein were separated in Laemmli buffer on 8%–12% (v/v) SDS-PAGE gel (Thermo Fischer Scientific, Milan, Italy) and transferred to a polyvinylidene difluoride (PVDF) membrane using the iBlot 2 System (Thermo Fischer Scientific, Milan, Italy). Following 5-min incubation with EveryBlot Blocking Buffer (Bio-Rad, Milan, Italy), the membranes were probed overnight at 4 °C with anti-CD81 (sc-166029, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-COX-2 (#4842S, Cell Signaling Technology, Danvers, MA, USA), and anti-tubulin (#3873, Cell Signaling Technology, Danvers, MA, USA) antibodies. Membranes were then incubated for 1-h RT with goat anti-mouse IgG Alexa Fluor 680 antibody or goat anti-rabbit IgG Alexa Flour 750 antibody (Thermo Fisher Scientific, Milan, Italy) and visualized at the Odyssey Infrared Imaging System (LI-COR® Bioscience, Lincoln, NE, USA).
The IL-1β, IL-6, and TNF-α concentrations were measured in 100 μL of culture medium from RAW 264.7 cells treated with EV (with or without anti-miR-214) or 10 ng/mL LPS (CM) by mouse uncoated ELISA kits (Thermo Fisher Scientific, Milan, Italy) according to manufacturer’s instructions. The absorbance was measured at 450 nm at a microplate reader (BioTek, Winooski, VT, USA). The amount of each cytokine in the media was interpolated within the standard curves (GraphPad Prism 7 software). The results were normalized to the number of cells.
A Griess reaction was used to measure the NO concentration in the culture medium of RAW 264.7 macrophages. Briefly, 100 μL of the culture medium conditioned by RAW 264.7 cells treated with EV (with or without anti-miR-214) or 10 ng/mL LPS was mixed 1:1 with the Griess reagent (1% sulfanilamide, 0.1% N-1-naphthalenediamine dihydrochloride, and 2.5% H3PO4, Merk Life Sciences S.r.l., Milan, Italy) and transferred to 96-well plates. After 10 min of incubation at RT, the absorbance was measured at 540 nm with a microplate reader (BioTek, Winooski, VT, USA). A sodium nitrite (NaNO2) standard curve was used to calculate the amount of nitrite in each sample. The results were normalized to the number of cells.
The Lactate Colorimetric Assay Kit (BioVision, purchased from Vinci-Biochem, Florence, Italy) was used according to the manufacturer’s instructions to measure lactate production in the conditioned media (CM) of RAW 264.7 macrophages following 24-h treatment with EV. Data normalization was obtained by directly counting the number of cells to get a final result of lactate production (nM) by 1 × 105 cells.
A number of (5 × 104 RAW 264.7) cells treated for 24 h with EV were seeded onto Seahorse XFe96 microplates and evaluated with the Seahorse XFe96 Extracellular Flux Analyzer (Seahorse Bioscience, Billerica, MA, USA) for their glycolytic metabolism using the Glycolytic Rate Assay Kit (Agilent Technologies, Santa Clara, CA, USA), according to the manufacturer’s instructions. All the experiments were performed at 37 °C and normalized via cell protein measure with a Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Milan, Italy). The Seahorse XF Report Generator automatically calculated the parameters from wave data that were exported to GraphPad Prism software.
Cells were harvested by Accutase (Euroclone, Pero, Italy), collected in flow cytometer tubes (2 × 105 cells/tube), and stained for 1 h at 4 °C with anti-VE-cadherin F-8 conjugated with Alexa Fluor 488 (sc-9989 AF488; Santa Cruz Biotechnology, Santa Cruz, CA, USA). Following a wash in PBS, the cells were analyzed at BD FACSCanto II (BD Biosciences, Milan, Italy), calibrated by using cells incubated with Alexa Fluor 488-conjugated irrelevant IgG, and 1 × 104 events/sample were analyzed.
Millicell Cell Culture Inserts (Merck Life Sciences S.r.l., Milan, Italy) were placed onto 24-weel plates and polycarbonate filters coated with 0.25 µg/µL Matrigel. Some (1 × 105) murine endothelial cells were cultured to confluency and treated for 24 h with the CM of RAW 264.7 previously subjected to EV treatment. Then, the permeability treatment was removed and albumin–fluorescein isothiocyanate conjugate (BSA-FITC, Merk Life Sciences, Milan, Italy) was incubated for 60 min. Some (100 μL) of the medium was collected from the receiver tray, transferred to a 96-well plate, and read on the Fluoroskan Ascent FL fluorescent plate reader (ThermoFiher Scientific, Milan, Italy) at 485 nm (excitation) and 535 nm (emission).
The trans-endothelial migration ability of B16 melanoma cells was evaluated towards murine endothelial cells monolayers pretreated for 24 h with CM of RAW 264.7 macrophages, previously receiving melanoma-derived EV. Briefly, Millicell Cell Culture Inserts (Merck Life Sciences S.r.l., Milan, Italy) were placed onto 24-weel plates, and the polycarbonate filters with 8 µm-diameter pores were coated with 0.25 µg/µL Matrigel. Some (1.0 × 105) endothelial cells were cultured to confluency in the upper compartment. The day before the experiment, B16 cells were labeled with carboxyfluorescein diacetate succinimidyl ester (CFDA-SE, Thermo Fisher Scientific, Milan, Italy). Some (2.5 × 104) CFDA-SE labeled B16 cells were seeded into the upper chamber and allowed to migrate for 6 h without any FBS gradient. Migrated cells were fixed in methanol for 1 h at 4 °C and observed at the SP8 confocal microscope (Leica Microsystems, Milan, Italy).
All data were obtained based on at least three independent experiments. Statistical analysis was performed with GraphPad Prism 6 software by t-test, one-way analysis of variance (ANOVA), and two-way ANOVA, as specified in each figure legend. Values are presented as mean ± standard deviation (SD). The p-values are presented as * p < 0.05, ** p < 0.01, and *** p < 0.001. Values are presented as the mean of independent experiments ± SD.
EV derived from B16, B16-LU, and B16-LI melanoma cells were evaluated by TEM (Figure 1A), evidencing a mixture of exosomes and microvesicles ranging from 50 to 400 nm in diameter, characterized by the expression of CD81 (Figure 1B; whole western blot membrane in Figure S1), one of the proper surface markers of EV. NTA revealed EV mixtures comprising exosomes of ≤100 nm in diameter and microvesicles of around 100–400 nm in length (Figure 1C). The mean values of the EV mixture released by each cell line under standard and chronic extracellular acidosis are 185.8 +/− 3.4 nm for control B16-derived EV vs. 174.4 +/− 2.5 nm for acid B16-derived EV, 150.9 +/− 14.3 nm for control B16-LU-derived EV vs. 208.6 +/− 7.9 nm for acid B16-LU-derived EV, and 185.5 +/− 10.1 nm for control B16-LI-derived EV vs. 142.9 +/− 3.2 nm for acid B16-LI-derived EV. Notably, for all the three cell lines, NTA highlighted a significant increase in EV release under extracellular acidosis compared to standard pH conditions, accounting for a ~five-fold increment in B16 cells, a ~four-fold increment in B16-LU cells, and a ~three-fold increment in B16-LI cells. Under standard pH conditions, no significant differences in the number of released EV were observed between the B16 and B16-LU cells, while B16-LI produced about a double amount of EV compared to B16 cells. Concerning acid EV, no significant differences were observed in the concentration of EV released by the three cell lines (Figure 1C, lower). Extracellular acidosis, besides affecting the amount of EV produced by each cell line, was found to also influence their cargos. In particular, the level of miR-214, a key “melano-miR” involved in melanoma progression and, conveyed in acid EV, was significantly higher than in control EV: miR-214-3p was increased ~2.5-fold in B16 and B16-LU acid-EV and ~4-fold in B16-LI acid-EV compared to control; miR-214-5p, despite less represented than the 3p strand, was upregulated in acid-EV of ~four-fold in all B16, B16-LU, and B16-LI-derived EV (Figure 1D). The increased miR-214 level in acid-EV was also observed in the A375, WM266-4, and Sk-Mel-2 human melanoma cell lines (Supplementary Figure S2).
There is increasing evidence that EV play a critical role in reprogramming not only tumor cells but also innate and adaptive immune cells of TME, among which macrophages are particularly abundant and present at all stages of tumor progression, playing either tumor suppression or tumor potentiation roles. We found that the uptake of melanoma-derived acid EV by RAW 264.7 macrophages induced their activation towards a pro-inflammatory phenotype. The nuclear translocation of the NF-ĸB transcription factor was observed in RAW 264.7 macrophages upon the uptake of EV derived from acid-adapted B16, B16-LU, and B16-LI cells. This translocation, on the contrary, was not visible when treated with control EV. Nuclear NF-ĸB staining was indeed ~1.4-fold higher in macrophages treated with acid EV compared to the ones treated with those released by melanoma cells under standard pH conditions and to the untreated RAW 264.7 macrophages. A slight but significant increase in nuclear NF-ĸB localization was also observed following the uptake of control B16-LU-derived EV compared to the untreated macrophages (Figure 2A). Moreover, we observed a moderately increased expression of COX-2 in RAW 264.7 macrophages treated with acid melanoma-derived EV compared to those treated with control EV (Figure 2B; whole western blot membranes in Figure S1). The mRNA expression level of inflammatory cytokines such as IL-1β, IL-6, and TNF-α confirmed the macrophage activation state upon the treatment with the acid EV derived from all the three melanoma cell lines used. IL-1β and IL-6 mRNA were upregulated ~1.4-fold and ~2-fold, respectively, in RAW 264.7 macrophages treated with acid EV compared to control EV from B16, B16-LU, and B16-LI cells; TNF-α mRNA expression level was ~1.4-fold higher in macrophages treated with acid-EV than in those treated with control-EV from B16 and B16-LU cells, reaching a ~4-fold increase with acid EV produced by B16-LI melanoma cells. Compared to untreated RAW 264.7 macrophages, even the treatment with control EV from B16, B16-LU, and B16-LI cells stimulates the expression of IL-1β, IL-6, and TNF-α, at least at the mRNA level (Figure 2C). The production of these cytokines was then evaluated by ELISA, revealing that the treatment with control EV did not change IL-1β, IL-6, and TNF-α secretion by RAW 264.7 macrophages compared to their untreated condition but confirming a significantly increased production when subjected to the treatment with acid EV derived from the three melanoma cell lines. In other words, EV from acid-adapted B16 cells induce a higher production of IL-1β, IL-6, and TNF-α compared to EV derived from B16 cells grown at the control pH levels. Likewise, the acid EV derived from the two metastatic cell lines B16-LU and B16-LI enhanced the production of these cytokines compared to the control EV (Figure 2D). In parallel with the enhanced cytokine production, RAW 264.7 macrophages also produced an increased amount of NO when treated with acid EV released by B16 (~2.5-fold), B16-LU (~1.5-fold), and B16-LI (~2-fold) cells compared to control EV. The treatment with EV derived from B16 and B16-LU cells grown at standard pH did not alter NO production compared to the untreated RAW 264.7 macrophages, while those released by standard pH B16-LI cells promote a ~four-fold-increased production (Figure 2E). Macrophages adjust their cellular phenotype upon their activation also reprogramming their metabolism. Indeed, it is known that macrophage activation is accompanied by metabolic reprogramming towards glycolysis [17]. Following the uptake of acid EV produced by acid-adapted B16, B16-LU, and B16-LI cells, macrophages showed a boosted glycolytic metabolism when compared to macrophages treated with control EV, as witnessed by the doubled lactate production (untreated macrophages showed the lowest lactate production among all the conditions assessed, nearby the value of 0.2 nM, normalized on the number of producing cells) (Figure 2F). A deeper metabolic analysis performed with the Seahorse XFe96 analyzer using the Glycolytic Rate Assay kit revealed increased basal glycolysis coupled with a higher proton efflux rate in RAW 264.7 macrophages treated with acid EV compared to those receiving control EV (Figure 2G). These data confirmed that acid EV-induced macrophage activation is paralleled by a metabolic switch toward glycolysis. This last observation contributes to the characterization of the activated phenotype of acid EV-treated macrophages, in turn likely suggesting a tool to maintain a pro-tumoral extracellular acidosis. LPS treatment was used as a positive control of macrophage activation: RAW 264.7 cells treated with LPS showed a doubled increase in NF-ĸB nuclear staining (Supplementary Figure S3A), a dramatic overproduction of NO (Supplementary Figure S3B), and a boosted glycolysis as lactate production increases eight times compared to control/untreated cells (Supplementary Figure S3C) going in parallel with augmented basal glycolysis and an increased proton efflux rate (Supplementary Figure S3D).
By using anti-miRNA technology to inhibit miR-214 activities, we observed a reversion of the inflammatory state acquired by macrophages following the treatment with EV released by acid-adapted melanoma cells conveying high levels of miR-214. We firstly observed a diminished mRNA expression of IL-1β, IL-6, and TNF-α. In detail, the anti-miR-214, given together with acid EV from B16 cells, induced a slight decrease in IL-1β and IL-6 (even though not significant) and a ~2.5-fold decrease in TNF-α expression; when administered with B16-LU-derived and B16-LI-derived acid EV, the anti-miR-214 significantly decreased IL-1β, IL-6, and TNF-α (Figure 3A). By ELISA, we confirmed the impairment of the release of these cytokines, in particular, RAW 264.7 macrophages treated with B16-derived acid EV and anti-miR-214 reduced the production of IL-1β, IL-6, and TNF-α; following the treatment with acid EV derived from B16-LU cells in the presence of anti-miR-214, RAW 264.7 macrophages underwent a decreased production of IL-1β while IL-6 appeared unaffected; TNF-α production was instead reduced, although not significantly; RAW 264.7 macrophages treated with acid EV from B16-LI cells in the presence of anti-miR-214 showed a mild but significant decrease in the secretion of IL-1β and IL-6 and a ~1.5 fold--reduced production of TNF-α (Figure 3B). In line with these observations, NO production was significantly impaired by anti-miR-214 when administered together with acid EV from B16 or B16-LI cells, while no significance was obtained with acid EV from B16-LU cells, although a trend to decrease was visible (Figure 3C).
To strengthen the results obtained by anti-miRNA treatment, we adopted an opposite strategy determining whether a forced miR-214 expression in RAW 264.7 macrophages correlates with higher levels of proinflammatory activation. Once obtained miR-214-over-expressing (miR-214+) cells by genetic selection, we enriched the high-expressing population by cell sorting for GFP positivity. The quality of the procedure was verified by the real-time qPCR assessment of miR-214-3p/5p expression in the control and miR-214+ RAW 264.7 macrophages, showing a four-fold and a six-fold increase in 3p and 5p strand, respectively (Figure 4A). miR-214+ cells produced a higher amount of NO compared to the controls (Figure 4B), also displaying a ~2-fold increase in IL-1β and IL-6 mRNA expression and a ~1.4-fold increase in the TNF-α mRNA levels (Figure 4C). ELISA confirmed the overproduction of IL-6 and TNF-α in miR-214+ macrophages compared to control, while no significant variation was observed in IL-1β release (Figure 4D).
Inflammation is a strong inducer of vascular permeability, so we questioned whether the macrophage activation induced by melanoma-derived acid EV enriched in miR-214 can promote tumor progression by altering the integrity of the endothelial cell layer. This would facilitate the trans-endothelial migration of melanoma cells, increasing the tumor intravasation rate. To verify that, we let RAW 264.7 macrophages treated with melanoma-derived control and acid EV condition their culture media for 24 h, and then, we administered the collected CM to murine endothelial cells to evaluate any variation in vascular permeability. Endothelial cells receiving CM of macrophages treated with melanoma-derived acid EV, in contrast to control EV, showed an impaired plasma membrane expression of VE-cadherin, the most important adhesion molecule for the stability of endothelial intercellular junctions [18] (Figure 5A). This observation was further strengthened by the permeability assay data obtained showing that a significantly increased amount of BSA-FITC passes through the endothelial cell monolayer incubated with CM of RAW 264.7 macrophages treated with acid EV from B16, B16-LU, or B16-LI cells compared to the control EV (Figure 5B). Notably, such an increased vascular permeability facilitated B16 melanoma cell trans-endothelial migration (Figure 5C). These data suggest that the acid-adapted melanoma cells via the miR-214-enriched-EV release and subsequent macrophage activation orchestrate the establishment of an inflamed microenvironment that, in turn, disrupts the vascular integrity facilitating tumor cell intravasation, the first key step of the metastatic dissemination.
The high proliferative rate of cancer cells is preferentially sustained by a glycolytic metabolism even in the presence of sufficient oxygen to guarantee phosphorylative oxidation, the so-called “Warburg effect” [19]. Such a boosted glycolysis leads to an excessive release of lactic acid in the extracellular milieu, inevitably causing the acidification of TME), further strengthened by the reduced lymphatic circulation and the high interstitial pressure typical of cancer tissues [20]. Thereby, almost all solid tumors experience an acidic extracellular pH ranging from 6.4 to 7.0 [21]. Huge evidence highlights that melanoma progression and metastatic dissemination are strongly promoted by the acidic TME. In the last decade, we and other research groups contributed to defining some of the tumor aggressive features fostered by extracellular acidosis, suggesting that the acidic TME affects each step of the metastatic cascade [22,23,24,25]. TME acidosis is known to participate actively in tumor progression and metastatic disease. Huge evidence highlights indeed that extracellular acidosis represents a perfect storm for metastasis development and it is now considered a new hallmark of cancer [25]. We previously reported that extracellular acidosis induces the acquisition of an extremely aggressive stem-like tumor phenotype endowed with a high ability to invade surrounding tissues, resist therapies, induce tumor vascularization, survive in the blood and lymphatic circulation, evade immune surveillance, and colonize secondary organs [15,23,24,26,27]. Notably, we showed that acid-adapted melanoma cells are less able to colonize secondary organs when injected alone into the mice’s tail vein but are capable to boost secondary organ colonization by the non-acid tumor cells when the co-injection of the two tumor sub-population was performed [28]. Such a contribution may likely rely on paracrine signals released by acid-adapted melanoma cells and received by the non-acid counterpart. In this context, EV may exert a crucial role. EV are emerging as critical messengers in cancer progression and metastasis [29]. By circulating through both blood and lymphatic vessels, EV create new hospitable sites for tumor growth, modulating inflammation, stromal organization, angiogenesis, coagulation, immune response, vascular permeability, and organotropism [7]. We observed that the release of EV by melanoma cells is strongly induced under extracellular acidosis, being acid-adapted melanoma cells capable to secrete a significantly higher number of EV compared to those cultured under standard pH conditions. This data is in line with previous findings showing that extracellular acidosis not only enhances the release of exosomes by tumor cells but also promotes significant modifications in the lipid component of the same particles [30,31,32]. Our data show that acid EV, compared to control EV, carry a high amount of miR-214, an onco-miRNA involved in the progression of different tumor types [13,33,34,35]. Focusing on melanoma, previous findings reported that miR-214 promotes the resistance to anoikis and the extravasation phases of the metastatic cascade [10,11,12]. In this study, we reveal a new possible way—promoted by the acidic TME—by which miR-214 can foster tumor trans-endothelial migration, acting not directly on melanoma cells but affecting macrophages. Macrophages are considered the principal regulators of several diseases, including cancer, and their accumulation in the tumor microenvironment is considered one of the hallmarks of cancer, also due to their plasticity to either reduce or potentiate tumor growth generating different local effectors. Indeed, we observed that miR-214 carried in EV released by acid-adapted melanoma cells is capable to induce a proinflammatory response in macrophages: upon NF-ĸB nuclear translocation together, macrophages increased the expression of COX-2 and the release of a high amount of IL-1β, IL-6, TNF-α, and NO, establishing an inflammatory microenvironment that in turn enhances vascular permeability facilitating melanoma cell trans-endothelial migration. In addition to NF-ĸB, COX-2, IL-1β, IL-6, TNF-α, and NO considered as markers of the activation phenotype expressed by macrophages exposed to miR-214-enriched EV, we also evaluated their metabolism. Generally, proinflammatory macrophages mainly rely on glycolysis and exhibit impaired tricarboxylic acid cycle and mitochondrial phosphorylative oxidation [36]. In our experimental model, we indeed observed that the proinflammatory macrophage activation induced by miR-214-enriched acid EV is coupled with a boosted glycolytic metabolism. The switch toward a glycolytic metabolism characterizing activated macrophages also offers the possibility to suggest a potential loop in which extracellular acidosis promotes the release of EV able to reprogram macrophage phenotype toward a prominent glycolytic activity sustaining, in turn, the acidosis of extracellular milieu. The use of anti-miRNA oligonucleotides, on one hand, and of miR-214-overexpression, on the other, to interfere with or induce, respectively, miR-214-related changes allowed us to verify that miR-214 is a key mediator of the inflammatory macrophage response stimulated by acid-adapted melanoma cells through EV. Despite the roles of miR-214 in the regulation of inflammation being much less studied, there is evidence suggesting its involvement in inflammatory responses. Indeed, miR-214 was found upregulated in THP-1 cells and human monocytes treated with proinflammatory advanced glycation end products (AGEs) [37] and in activated T cells, playing a critical role in T-cell proliferation and excitation through targeting phosphatase and tension homolog deleted on chromosome 10 (PTEN) [38], known to have remarkable anti-inflammatory activities [39,40]. miR-214 overexpression was also found to induce an inflammatory response in the ulcerative colitis experimental model by promoting NF-kB phosphorylation and subsequently IL6 expression through PTEN and PDZ-LIM domain-containing protein 2 (PDLIM2) inhibition [41]. More recently, Zhao Li and colleagues disclosed the existence of a mutual suppression feedback loop between miR-214 and adenosine A2A receptor (A2AR) signaling in the inflammatory response: briefly, they found that miR-214 overexpression promotes the release of inflammatory cytokines and, by directly binding to the 3′-UTR, downregulates A2AR expression; conversely, A2AR activity represses miR-214 expression in a PKA-NF-κB signaling-dependent manner. They also observed that NF-κB directly binds to the miR-214 promoter, resulting in transcriptional upregulation of miR-214 [42]. An in silico analysis by TargetScanHuman also suggests that, among putative miR-214 targets, besides PTEN, there are negative regulators of NF-ĸB, such as the NFKB inhibitor interacting Ras-like 2 (NKIRAS2) [43] and the NF-κB-repressing factor (NKRF), that specifically counteracts several basal NF-κB activities [44]. Thus, miR-214 represents a potential novel regulator of inflammation, despite the fact that its mechanisms of action in inflammation are not completely elucidated.
Overall, we revealed that the acidic TME potentiates the release by melanoma cells of miR-214-enriched EV that, in turn, promote the establishment of a macrophage-dependent inflammatory microenvironment facilitating vascular permeability and tumor trans-endothelial migration. Aware of the importance of translating this study in vivo, we plan to validate our results by using a model of experimental metastasis in syngeneic mice. In line with our observations, very recently, Bychkov and colleagues demonstrated that extracellular acidosis, via inducing metastatic melanoma cells to release EV with advanced pro-tumoral activity, sustained the acquisition of an increased malignant phenotype of melanoma cells themselves, also redirecting keratinocytes towards pro-tumoral activities [45]. This preclinical in vitro study allowed us to identify another mechanism through which the acidic TME may foster tumor progression and dissemination, exploiting the great potential of EV to convey pro-tumoral molecular signals capable of altering the microenvironment of secondary locations and prepare for metastatic colonization. These data further strengthen the awareness of the pro-tumoral activity of the acidic TME, suggesting that miR-214-enriched EV potentiate melanoma intravasation. | true | true | true |
|
PMC9600018 | Sen Ren,Jing Chen,Jiahe Guo,Yutian Liu,Hewei Xiong,Boping Jing,Xiaofan Yang,Gongchi Li,Yu Kang,Cheng Wang,Xiang Xu,Zhenyu Liu,Maojie Zhang,Kaituo Xiang,Chengcheng Li,Qianyun Li,Hans-Günther Machens,Zhenbing Chen | Exosomes from Adipose Stem Cells Promote Diabetic Wound Healing through the eHSP90/LRP1/AKT Axis | 14-10-2022 | adipose-derived stem cell,exosomes,diabetic wound,oxidative stress,heat shock protein 90 | Oxidative damage is a critical cause of diabetic wounds. Exosomes from various stem cells could promote wound repair. Here, we investigated the potential mechanism by which exosomes from adipose-derived stem cells (ADSC-EXOs) promote diabetic wound healing through the modulation of oxidative stress. We found that ADSC-EXOs could promote proliferation, migration, and angiogenesis in keratinocytes, fibroblasts, and endothelial cells. Furthermore, ADSC-EXOs reduced the reactive oxygen species (ROS) levels in these cells and protected them against hypoxic and oxidative stress damage. Finally, the local injection of ADSC-EXOs at wound sites significantly increased collagen deposition and neovascularization while reducing ROS levels and cell death; thus, it led to accelerated diabetic wound closure. The mechanism underlying ADSC-EXO functions involved heat-shock protein 90 (HSP90) expressed on the cell surface; these functions could be inhibited by an anti-HSP90 antibody. Exosomal HSP90 could bind to the low-density lipoprotein receptor-related protein 1 (LRP1) receptor on the recipient cell membrane, leading to activation of the downstream AKT signaling pathway. Knockdown of LRP1 and inhibition of the AKT signaling pathway by LY294002 in fibroblasts was sufficient to impair the beneficial effect of ADSC-EXOs. In summary, ADSC-EXOs significantly accelerated diabetic wound closure through an exosomal HSP90/LRP1/AKT signaling pathway. | Exosomes from Adipose Stem Cells Promote Diabetic Wound Healing through the eHSP90/LRP1/AKT Axis
Oxidative damage is a critical cause of diabetic wounds. Exosomes from various stem cells could promote wound repair. Here, we investigated the potential mechanism by which exosomes from adipose-derived stem cells (ADSC-EXOs) promote diabetic wound healing through the modulation of oxidative stress. We found that ADSC-EXOs could promote proliferation, migration, and angiogenesis in keratinocytes, fibroblasts, and endothelial cells. Furthermore, ADSC-EXOs reduced the reactive oxygen species (ROS) levels in these cells and protected them against hypoxic and oxidative stress damage. Finally, the local injection of ADSC-EXOs at wound sites significantly increased collagen deposition and neovascularization while reducing ROS levels and cell death; thus, it led to accelerated diabetic wound closure. The mechanism underlying ADSC-EXO functions involved heat-shock protein 90 (HSP90) expressed on the cell surface; these functions could be inhibited by an anti-HSP90 antibody. Exosomal HSP90 could bind to the low-density lipoprotein receptor-related protein 1 (LRP1) receptor on the recipient cell membrane, leading to activation of the downstream AKT signaling pathway. Knockdown of LRP1 and inhibition of the AKT signaling pathway by LY294002 in fibroblasts was sufficient to impair the beneficial effect of ADSC-EXOs. In summary, ADSC-EXOs significantly accelerated diabetic wound closure through an exosomal HSP90/LRP1/AKT signaling pathway.
Diabetic wounds are common complications of diabetes mellitus; such wounds pose a considerable clinical challenge [1]. Because delayed or non-healing wounds can lead to amputation and mortality, they constitute a heavy burden on patients’ families and society [2]. Currently, there are no effective treatment methods for such wounds. Oxidative stress damage, mediated by excessive production of reactive oxygen species, is a critical obstacle to healing in diabetic wounds [2,3]. Under diabetic conditions, local hypoxia and high glucose levels cause mitochondria to produce large amounts of ROS [4]. The presence of high ROS concentrations leads to sustained pro-inflammatory cytokine secretion and matrix metalloproteinase overproduction [5]. Furthermore, excessive ROS can impair skin cells (e.g., endothelial cells, keratinocytes, and fibroblasts), resulting in restricted neovascularization, decreased granulation tissue formation, and extracellular matrix deposition [6]. Previous studies have demonstrated that mesenchymal stem cells (MSCs) have therapeutic effects on diabetic wound healing [7]. However, the clinical implementation of MSC-based therapies has been limited by challenges such as immune rejection, risk of tumorigenesis, low transplantation efficacy, and ethical conflicts [8]. There is increasing evidence that the beneficial effects of MSCs mainly rely on their paracrine factors (e.g., growth factors, cytokines, and extracellular vesicles) [9]. Exosomes are small extracellular vesicles (30–150 nm in diameter) that participate in intercellular communication by delivering proteins, lipids, DNA, and RNA [10]. Recent studies have shown that exosomes from different MSCs can accelerate diabetic wound healing [11,12,13]. Adipose-derived mesenchymal stem cells (ADSCs) are regarded as highly promising stem cells resources because of their easy harvest, abundant content, and low immunogenicity characteristics [14]. Recent studies have also shown that ADSC-EXOs can promote skin wound healing through various mechanisms [15,16,17]. However, the specific effects of ADSC-EXOs on diabetic wound healing and the underlying mechanism require further investigation. The heat-shock protein (HSP) family consists of a group of highly conserved proteins that respond to multiple stresses (e.g., heat, hypoxia, trauma, and starvation) [18]. Most HSPs act as molecular chaperones to assist protein folding and repair damaged proteins [19]. Initial investigations suggested that HSPs only functioned intracellularly. However, more recent studies have revealed that HSPs can be secreted into the external environment to mediate cellular communication; they have important roles in skin wound healing [20]. Defects in HSP function associated with diabetes might contribute to the delayed healing in diabetic wounds. Uncontrolled oxidative stress is a typical component of diabetes-related complications. Thus, there is a need to explore the antioxidant ability of HSPs for the management of diabetic wounds. Previous studies have shown that extracellular HSP90 (eHSP90) participates in normal wound healing; exogenous administration of eHSP90 could promote skin cell migration, thus accelerating the closure of normal and diabetic wounds [21,22,23]. The functions of extracellular HSPs depend on their interactions with cellular receptors such as Toll-like receptors 2 and 4 (TLR2 and TLR4) and low-density lipoprotein receptor-related protein 1 (LRP1) [24]. Because they lack signal peptides, HSPs are secreted into the extracellular environment mainly through exosomes [21,25]. Therefore, eHSP90 delivery through ADSC-EXOs could have therapeutic benefits. In the present study, we hypothesized that ADSC-EXOs communicate with skin cells to provide signals that protect the skin cells against oxidative stress damage. We demonstrated that ADSC-EXOs could reduce intracellular ROS levels in skin cells while protecting skin cells from excessive ROS- and hypoxia-induced cell death both in vivo and in vitro. These protection and antioxidant effects were mediated by HSP90 present on the exosomal surface, which bound to LRP1 and activated the downstream AKT signaling pathway. Our findings provide important insights for therapeutic management of diabetic wound healing.
Human umbilical vein endothelial cells (HUVECs; #GDC166) and HaCaT cells (#GDC106) were purchased from the China Center of Type Culture Collection (CCTCC, Wuhan, China) and cultured in accordance with the supplier’s instructions. The inhibitors of PI3K (LY294002) and ERK1/2 (U0126) were purchased from MedChemExpress (Monmouth Junction, NJ, USA; HY-10108 and HY-12031A). Normal mouse IgG and anti-HSP90 mouse monoclonal IgG were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). The following primary antibodies were utilized: anti-AKT (4691S), anti-p-AKT (4060S), anti-ERK (4695S), anti-p-ERK (4370S), anti-NF-κb (8242T), and anti-p-NF-κb (3033T) (all from Cell Signaling Technology, Danvers, MA, USA); anti-LRP1 (ab92544), anti-CD9 (ab236630), anti-CD63 (ab134045), and anti-GAPDH (ab8245) (all from Abcam, Waltham, MA, USA); anti-GM130 (11308) and anti-ꞵ-tubulin (66240) (both from Proteintech, Wuhan, China); and anti-HSP70 (sc-32239) and anti-HSP90 (sc-13119) (both from Santa Cruz, Dallas, TX, USA).
Human adipose tissues and foreskins were obtained in accordance with procedures approved by the Ethics Committee at the Tongji Medical Collage of Huazhong University of Science and Technology. Primary fibroblasts were isolated from foreskins derived from routine pediatric circumcisions, using previously described protocols [26]. ADSCs were isolated and cultured in accordance with our established protocol [27]; ADSCs at passages 3–8 were used for experiments. Flow cytometry was conducted to identify ADSC phenotypes. ADSCs were incubated for 1 h with anti-CD34-BV421, anti-CD44-APC, anti-CD105-PE, anti-CD73-FITC, anti-CD31-FITC, and anti-CD90-FITC antibodies (all from Biolegend, San Diego, CA, USA); fluorescence was observed thereafter. For multi-lineage differentiation potential assays, ADSCs were incubated separately with adipogenic medium and osteogenic differentiation medium (Cyagen Biosciences Inc., Guangzhou, China). Cells were stained with Oil Red O and Alizarin Red, respectively. Subsequently, the stained cells were imaged using a microscope.
ADSC-EXOs were isolated from the supernatant of adipose stem cells by differential centrifugation as described previously [28]. Briefly, the culture media were centrifuged at 1000× g for 10 min, 3000× g for 25 min, and 13,000× g for 40 min to eliminate dead cells, debris, and large macrovesicles separately. Then, the supernatants were centrifuged at 120,000× g for 70 min twice to obtain the exosomes pellets. Finally, the exosomes pellets were resuspended in phosphate-buffered saline (PBS) and stored at −80 °C. The protein concentration was determined by a BCA protein assay kit (Beyotime Biotechnology, Shanghai, China). The morphology of ADSC-EXOs was captured by transmission electron microscope (Hitachi, Tokyo, Japan). The size and concentration of the ADSC-EXOs was evaluated by nanoparticle tracking analysis (Wayen Biotechnologies, Shanghai, China). The exosomal-positive markers CD63, CD9, HSP70, and -negative markers GM130 and β-tubulin were detected by Western blot.
ADSC-EXOs were labeled with PKH26 fluorescent dye (Sigma-Aldrich, St. Louis, MO, USA) in accordance with the manufacturer’s instructions. Cells were seeded in confocal dishes and incubated with PKH26-labeled ADSC-EXOs for various intervals. Phalloidin (Yeasen Biotech Co., Shanghai, China) and DAPI (Solarbio, Beijing, China) were used to label the cytoskeleton and nucleus before image acquisition. Images were acquired using a confocal microscope (Nikon, Tokyo, Japan).
Fibroblasts, HUVECs, and HaCaT cells were treated with phosphate-buffered saline (PBS) or ADSC-EXOs (20 µg/mL) for 48 h; cell proliferation was evaluated by EdU kits (Beyotime Biotechnology). The migration abilities of fibroblasts, HUVECs, and HaCaT cells were assayed using 24-well Transwell Chambers (Corning Inc., Corning, NY, USA). In brief, cells suspended in Dulbecco’s modified Eagle medium without serum were added to the upper chamber and treated with PBS or ADSC-EXOs (5 µg/mL and 10 µg/mL). The lower chamber was filled with 600 µL of complete culture media consisting of 10% fetal bovine serum. After incubation at 37 °C for 24 h, cells that had migrated to the bottom surface were stained with crystal violet (Solarbio, Beijing, China). For the tube formation assay, HUVECs were seeded in 48-well plates that had been precoated with Matrigel Basement Membrane Matrix (BD Biosciences, NJ, USA). After incubation with PBS and ADSC-EXOs (10 µg/mL or 20 µg/mL) for 2 h, 4 h, and 8 h, tube formation was imaged by microscopy. The total number of loops, nodes, and branch points was calculated and used to quantify the tubular networks.
Cellular models of oxidative stress and hypoxia were generated using H2O2 and cobalt chloride (CoCl2) (both from Sigma-Aldrich, St. Louis, MO, USA). HaCaT cells, fibroblasts, and HUVECs were seeded and treated with different concentrations of H2O2 for 6 h or CoCl2 for 24 h. Cell viabilities were analyzed using a CCK-8 kit (Dojindo, Kumamoto, Japan). The absorbance values at 450 nm were measured using a microplate reader (Tecan, Männedorf, Switzerland). Cell viabilities were also assessed by Calcein-AM/PI double staining (Yeasen Biotech Co., Shanghai China). Subsequently, the stained cells were imaged using a fluorescence microscope (Olympus, Tokyo, Japan).
Intracellular ROS levels in the skin cells of in vitro models were analyzed using a ROS detection kit (2,7-dichlorofluorescein diacetate, DCFH-DA) from the Nanjing Jiancheng Bioengineering Institute (Nanjing, China). After different treatments, cells were incubated with DCFH-DA for 30 min, in accordance with the manufacturer’s instructions. Then, cellular fluorescence was observed using a fluorescence microscope or a flow cytometer (BD Biosciences, NJ, USA).
After different treatments, fibroblasts were incubated with a Annexin V-FITC Apoptosis Detection Kit (Sigma-Aldrich, St. Louis, MO, USA), in accordance with the manufacturer’s instructions. The fluorescence of stained cells was detected by fluorescence-activated cell sorting using a flow cytometer (BD Biosciences, NJ, USA) and analyzed by FlowJo V10 software.
Equal exosomal proteins (20 µg) were separated on sodium dodecyl sulfate-polyacrylamide gel (SDS-PAGE) and subjected to silver staining (Beyotime Biotechnology, Shanghai, China). Then, the gels were cut into seven parts according to the protein molecular weight. Next, proteins in the stained bands were identified with the LC-MS/MS experiment (maXis impact UHR-TOF, Bruker, Germany). Raw data files were possessed by the Data Analysis software (Compass data analysis v4.1, Bruker, Germany). Protein identification was performed utilizing PEAKS 8.5 software (BSI, Canada). The GO categories (http://geneontology.org) were used for defining the cellular component, molecular function, and biological process of involved proteins.
IgG, anti-HSP90 antibody, and an equal volume of PBS were, respectively, mixed with ADSC-EXOs at a mass ratio of 10:1 and incubated at 37 °C for 30 min. The following are the three treatment groups used for subsequent experiments: EXOs groups, EXOs + IgG groups, and EXOs + anti-HSP90 groups.
For gene knockdown, sh-LRP1 and control lentiviruses were synthesized by Genechem Co., Ltd. (Shanghai, China). Cells were seeded into six-well plates for 1 day. The culture medium was then refreshed; cells were transfected with sh-LRP1 or control lentiviruses for 3 days. Gene-silencing efficiency was validated by Western blotting. The shRNA sequences are shown in Supplementary Table S1.
Equal amounts of total protein (20–40 μg) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride membranes. The membranes were blocked with 5% w/v bovine serum albumin for 1 h at room temperature, then incubated with primary antibodies overnight at 4 °C. Subsequently, membranes were incubated with secondary antibodies for 2 h and exposed to X-ray film for visualization by the BioSpectrum Imaging System (UVP, CA, USA).
All animal experiments were approved by the Animal Care Committee of Tongji Medical College. The effects of ADSC-EXOs on diabetic wound healing were evaluated in a diabetic mouse model with a cutaneous wound. Male C57BL/6 mice (20–25 g) were used in the present study. Diabetes was induced by intraperitoneal injection of streptozotocin (50 mg/kg) for 5 consecutive days; control mice received an equal volume of citrate buffer. Blood glucose was monitored for 2 weeks thereafter. Mice with blood glucose levels > 16.7 mM (300 mg/dL) were regarded as diabetic mice; they were monitored for an additional 4 weeks before the induction of full-thickness skin wounds. Mice were anesthetized by intraperitoneal injection of pentobarbital solution (50 mg/kg; Sigma-Aldrich). A full-thickness round skin wound (9 mm in diameter) was created on the back of each mouse; wound edges were sutured with silicone rings (1.0 cm in diameter) to prevent wound contraction. Then, the wounded mice were randomly divided into five treatment groups: Nor + PBS (normal mice treated with PBS); Dia + PBS (diabetic mice treated with PBS); Dia + EXOs (diabetic mice treated with EXOs); Dia + EXOs + IgG (diabetic mice treated with EXOs and IgG); and Dia + EXOs + anti-HSP90 (diabetic mice treated with EXOs and anti-HSP90 antibody). ADSC-EXOs (50 µg per mouse) and PBS (equal volume to the ADSC-EXOs suspension) were subcutaneously injected into five sites around the wounds once after the wounds were created. At days 0, 3, 8, and 13, the wounds were photographed and analyzed using ImageJ software (v1.46r).
Wound tissue samples were harvested on day 13 after surgery and fixed using 4% paraformaldehyde. After dehydration by ethanol, the tissues were embedded in paraffin and sliced into 5 μm thick longitudinal sections. The degree of wound re-epithelialization was analyzed using hematoxylin and eosin staining. The degree of collagen deposition was evaluated using Masson’s staining. To evaluate wound angiogenesis, immunofluorescence staining of α-SMA and CD31 was performed in wound tissue samples. To evaluate cell proliferation in wound tissue, we performed immunofluorescence staining of Ki67 and immunohistochemical staining of proliferating cell nuclear antigen (PCNA). Cell apoptosis in tissues was determined using a TUNEL fluorescence kit (BD Biosciences, NJ, USA). To detect ROS levels in wound sections, frozen wound tissue samples were cryosectioned; they were then stained with dihydroethidium at room temperature for 1 h. Images were captured with a fluorescence microscope and analyzed using ImageJ software (v1.46r).
All statistical data are shown as means ± standard errors of the mean; the data were evaluated using GraphPad Prism 7. Student’s t-test was applied to compare differences between two groups; one-way or two-way analysis of variance and Bonferroni post hoc analysis were used for comparisons of ≥3 groups. p-values < 0.05 were considered statistically significant (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001).
Primary ADSCs were successfully extracted from human adipose tissue and identified by their morphology, surface markers, and multi-lineage differentiation potential. Flow cytometry analysis showed that large proportions of ADSCs expressed CD73 (98.4%), CD90 (94.5%), CD44 (99.6%), and CD105 (99.2%); small proportions of ADSCs expressed CD34 (1.42%) and CD31 (1.24%) (Figure S1A). Microscopy analysis revealed that ADSCs had typical fibroblast-like morphology and grew in a swirled pattern (Figure S1B). The presence of intracellular lipid droplets and calcium nodules indicated that ADSCs could differentiate into adipocytes and osteoblasts (Figure S1C,D). Exosomes were isolated from ADSC culture supernatants by differential ultracentrifugation. Transmission electron microscopy analysis showed that ADSC-EXOs had cup-shaped morphology (Figure S2A,B). Western blotting analysis indicated that ADSC-EXOs expressed exosomal markers CD63, CD9, and HSP70; they did not express GM103 or β-tubulin (Figure S2C). Nanoparticle tracking analysis revealed that the mean diameter of ADSC-EXOs was 126.3 ± 0.1 nm (Figure S2D). These results confirmed that ADSC-EXOs had successfully been isolated.
New granulation tissue formation requires precise interactions among keratinocytes, fibroblasts, and endothelial cells [29]. First, we investigated the effects of ADSC-EXOs on proliferation in these cells (i.e., keratinocytes, fibroblasts, and endothelial cells) by EdU assays. The percentages of proliferating cells were significantly increased after ADSC-EXO treatment; fibroblasts demonstrated the greatest benefit (Figure 1A,B and Figure S3). Next, the effects of ADSC-EXOs on the migration of HaCaT cells, fibroblasts, and HUVECs were evaluated by Transwell assays. As expected, after treatment with 5 µg/mL and 10 µg/mL ADSC-EXOs, the numbers of migrated HUVECs increased by 1.65-fold and 3.65-fold, respectively, compared with untreated controls; the corresponding numbers of migrated fibroblasts increased by 3.08-fold and 4.52-fold; and the corresponding numbers of migrated HaCaT cells increased by 1.71-fold and 3.11-fold (Figure 1C,D and Figure S4). Tube formation assays were used to evaluate the ability of ADSC-EXOs to promote angiogenesis in HUVECs. The numbers of capillary-like tubular structures and junctions were significantly increased after 2 h, 4 h, and 8 h of incubation with ADSC-EXOs (Figure S5). Taken together, these data demonstrated that ADSC-EXOs could enhance proliferation, migration, and angiogenesis in skin cells during in vitro experiments.
We used H2O2 and CoCl2 to simulate an excess oxidative stress and hypoxia microenvironment in diabetic wounds. Flow cytometry analysis showed that H2O2 triggered a sharp increase in the intracellular ROS levels of HaCaT cells, fibroblasts, and HUVECs; this trend could be partially reversed by treatment with ADSC-EXOs (Figure S6A–F). DCFH fluorescence analysis also demonstrated that ADSC-EXO treatment could significantly attenuate H2O2-induced ROS overproduction in fibroblasts (Figure 1E,F). Furthermore, the proportion of apoptotic fibroblasts was increased after H2O2 exposure; ADSC-EXO treatment fully abolished this harmful effect, as demonstrated by flow cytometry analysis of Annexin V/PI (Figure 1G,H). CCK8 assays also revealed that ADSC-EXO treatment could partially protect skin cells against H2O2- and CoCl2-induced cell death; fewer than 65% of HUVECs survived after H2O2 or CoCl2 exposure, while ADSC-EXO treatment greatly increased the proportions of surviving HUVECs regardless of exposure (Figure S6G,H). The rate of HaCaT cell survival in the H2O2 model was increased from 73.8% to 96.0% by ADSC-EXO treatment; it was increased from 66.1% to 86.3% in the CoCl2 model (Figure S6I,J). Furthermore, the rate of fibroblast survival in the H2O2 model was increased from 12.5% to 74.4% by ADSC-EXO treatment; it was increased from 49.7% to 77.6% in the CoCl2 model (Figure 1I,J). In brief, ADSC-EXO treatment could reduce the intracellular ROS level in skin cells and protect those cells against damage from hypoxia and oxidative stress.
To characterize the fates of exosomes among recipient cells, we explored whether ADSC-EXOs could be internalized by skin cells. For this purpose, we labeled ADSC-EXOs with PKH26 and co-cultured them with 293T cells, HUVECs, HaCaT cells, or fibroblasts. Red fluorescence was clearly observed in the cytoplasm of 293T cells, HUVECs, and HaCaT cells; it was nearly absent from fibroblasts (Figure 2A). To confirm the reliability of the findings, fibroblasts were co-cultured with Schwann cells, which have been shown to internalize exosomes [28]; the co-cultured cells were then incubated with PKH26-labeled ADSC-EXOs for 2 h, 6 h, 12 h, or 24 h. Similarly, red fluorescence was observed in the cytoplasm of Schwann cells but was generally absent from fibroblasts (Figure 2B). We concluded that ADSC-EXOs were not easily internalized by fibroblasts. Thus, we presumed that the effect of ADSC-EXOs on fibroblasts mainly relied on ligand–receptor-induced intracellular signaling activation. This type of mechanism may also participate in the effects of ADSC-EXOs on keratinocytes and endothelial cells. To investigate the exosomal membrane proteins that mediate the effects of ADSC-EXOs, we used mass spectrometry to identify the protein profiles of ADSC-EXOs (Figure 2C). In total, 1106 proteins were identified; gene ontology analysis was conducted to classify functional categories of all identified proteins (Tables S2 and S3). Gene ontology analysis showed that 77 proteins were located in the cell surface category; these were presumed to localize at the exosomal surface. Next, the 10 most abundant surface proteins were identified; these included HSP90 (Figure 2D). Previous studies have shown that HSP90 is present on the surface of tumor cell-secreted exosomes and has important roles in tumor invasion and metastasis [24]. Western blotting analysis also confirmed the high abundance of HSP90 in ADSC-EXOs (Figure 2E). Taken together, these data suggested that HSP90 might be the main effector of ADSC-EXOs.
To determine whether eHSP90 is essential for ADSC-EXO function, we used an anti-HSP90 antibody to neutralize this exosomal surface protein. CCK8 assays showed that ADSC-EXO treatment could significantly decrease H2O2- and CoCl2-induced cell death; this protective effect was eliminated upon pretreatment with an anti-HSP90 antibody rather than an IgG-isotype control (Figure 3A,B). This result was confirmed by Calcein-AM/PI staining analysis (Figure 3C,D). Additionally, H2O2 exposure caused a 3.46-fold increase in the intracellular ROS level in fibroblasts compared with untreated control fibroblasts; the intracellular ROS level was decreased (1.73-fold lower than untreated control fibroblasts) by ADSC-EXO treatment but increased again (2.99-fold higher than untreated control fibroblasts) upon pretreatment with an anti-HSP90 antibody (Figure 3E,F). These results suggested that the anti-HSP90 antibody could abolish the ability of ADSC-EXOs to protect fibroblasts from oxidative stress damage; they were also compatible with the findings in our Transwell assays to characterize the promigratory effect of exosomal surface HSP90. In the Transwell assays, we found that the numbers of migrated fibroblasts increased by 9.08-fold and 8.85-fold after treatment with EXOs or EXOs + IgG; these large increases were significantly suppressed by pretreatment of EXOs with an anti-HSP90 antibody (Figure 3G,H). However, the number of migrated fibroblasts remained 1.94-fold higher in the EXOs + anti-HSP90 group than in the control group. These data confirmed that the anti-HSP90 antibody could antagonize the promigratory effect of ADSC-EXOs. In conclusion, eHSP90 contributed to the effects of ADSC-EXOs on cell migration and protection.
Western blotting analysis was conducted for further exploration of the intracellular signaling pathways activated in fibroblasts during incubation with ADSC-EXOs. We focused on ERK, AKT, and NF-κB pathways because they are reportedly activated by eHSP90 [22,30]. As shown in Figure 4A, ADSC-EXO treatment induced time-dependent phosphorylation of both AKT (p-AKT) and ERK (p-ERK); the maximum increases in p-AKT and p-ERK levels (2.40-fold and 2.03-fold, respectively) were achieved at 60 min after ADSC-EXO treatment; however, there was no increase in the phosphorylation of NF-κB (p-NF-κB). Furthermore, ADSC-EXO treatment also induced time-dependent phosphorylation of both AKT (p-AKT) and ERK (p-ERK) in HUVEC and HaCaT (Figure S7). Next, to clarify whether AKT and ERK signaling pathways were involved in the protective effects of ADSC-EXOs, fibroblasts were pre-treated with LY294002 (AKT inhibitor) and U0126 (ERK inhibitor). First, we confirmed that the ADSC-EXO-induced phosphorylation of AKT and ERK could be entirely inhibited by LY294002 and U0126, respectively; each type of phosphorylation could also be suppressed by the use of anti-HSP90 antibody (Figure 4B). Next, we used CCK8 assays to investigate the protective effects of ADSC-EXOs on fibroblasts exposed to H2O2 and CoCl2; these protective effects could be totally abolished by LY294002 (Figure 4C,D). This result was confirmed by Calcein-AM/PI staining analysis (Figure 4E,F). Similarly, DCFH fluorescence analysis demonstrated that LY294002 could attenuate the ability of ADSC-EXOs to reduce the intracellular ROS level (Figure 4G,H). Notably, U0126 could not antagonize the protective effect of ADSC-EXOs; it greatly promoted fibroblast survival under oxidative stress damage (Figure S8A). These data indicated that eHSP90 protected fibroblasts against hypoxic and oxidative stress damage through activation of the AKT signaling pathway. Transwell migration assays showed that the effect of ADSC-EXOs on promoting fibroblasts migration was significantly suppressed by LY294002; however, LY294002 alone had no effect on cell migration (Figure 4I,J). In contrast, U0126 alone promoted fibroblast migration but did not affect the promigratory function of EXOs (Figure S8B,C). These data confirmed that ADSC-EXOs promoted cell migration by activating the AKT signaling pathway rather than the ERK signaling pathway. Thus, we concluded that eHSP90 promoted cell migration and survival through activation of the AKT signaling pathway.
A previous study demonstrated that eHSP90 binding to LRP1 could promote skin cell migration [22]. To further investigate whether ADSC-EXO function was mediated by LRP1, we designed three pairs of shRNAs to inhibit LRP1 expression in fibroblasts. Of these pairs, sh-LRP1#3 showed the highest interference efficiency; it was used in subsequent experiments (Figure 5A). Western blotting analysis showed that ADSC-EXOs significantly increased the levels of p-AKT and p-ERK in control fibroblasts but not in LRP1 knockdown fibroblasts (Figure 5B). As expected, ADSC-EXOs exhibited a pro-survival effect against H2O2 and CoCl2 exposures in control fibroblasts; this beneficial effect was not observed in LRP1 knockdown fibroblasts (Figure 5C,D). This result was confirmed by Calcein-AM/PI staining analysis (Figure 5E,F). Dihydroethidium fluorescence analysis also demonstrated that ADSC-EXOs could partially reverse the H2O2-induced enhancement of intracellular ROS level in control fibroblasts but not in LRP1 knockdown fibroblasts (Figure 5G,H). Furthermore, the number of migrated cells was greatly increased by EXO treatment in control fibroblasts but only slightly increased in LRP1 knockdown fibroblasts (Figure 5I,J). These results demonstrated that the promigratory and protective effects of ADSC-EXOs were related to the presence of LRP1 on the recipient cell surface. Therefore, ADSC-EXOs activated the AKT signaling pathway in a LRP1-dependent manner.
To determine whether ADSC-EXOs could accelerate diabetic wound healing through interactions with surface protein HSP90, we established a full-thickness diabetic wound model. The wound-closure rate was more rapid in diabetic mice treated with EXOs than in untreated mice at day 8 after wounding; this EXO-induced rapid healing was significantly suppressed by an anti-HSP90 antibody (Figure 6A,B). Furthermore, wound healing was substantially slower in diabetic mice than in healthy mice. Next, we performed a histological analysis to assess the degrees of wound healing and regeneration. Hematoxylin and eosin staining analysis revealed that wound edges were significantly narrowed in the EXO-treated groups; anti-HSP90 antibody could completely inhibit this effect (Figure 6C,D). Furthermore, Masson’s trichrome staining analysis showed that collagen deposition was more extensive and better-organized in wounds treated with EXOs; this benefit was not observed in wounds treated with EXOs that had been incubated with an anti-HSP90 antibody (Figure 6E,F). Strikingly, ADSC-EXO treatment caused the levels of collagen deposition and wound edge contraction in diabetic mice to approach the levels found in healthy mice. These data suggested that ADSC-EXOs could promote collagen deposition and accelerate wound healing via eHSP90.
Diabetic wounds contain a microenvironment characterized by hypoxia and high oxidative stress, which results in cell dysfunction and the potential for apoptosis. The ROS levels in wound sections were detected by dihydroethidium staining. The red fluorescence signals were stronger in diabetic wounds than in non-diabetic wounds, indicating that diabetic wounds had high ROS levels; moreover, ROS levels in diabetic wounds were dramatically diminished after treatment with ADSC-EXOs, but this effect was partially reversed by the use of anti-HSP90 antibody (Figure 7A,B). These data indicated that ADSC-EXOs exhibited excellent ROS-scavenging performance. Consistent with these findings, TUNEL staining showed that there were more apoptotic cells in diabetic wounds than in non-diabetic wounds; the number of apoptotic cells was significantly reduced after treatment with ADSC-EXOs in diabetic wounds. Furthermore, the use of an anti-HSP90 antibody was sufficient to antagonize the anti-apoptosis effect of ADSC-EXOs (Figure 7C,D). Taken together, the above results indicated that eHSP90 has a vital role in the attenuation of oxidative stress-induced cell apoptosis. Immunofluorescence staining of CD31 and α-smooth muscle actin (α-SMA) on wound sections was performed to observe angiogenesis in vivo. The numbers of newly formed capillaries and new mature vessels were significantly higher in diabetic wounds treated with ADSC-EXOs than in untreated control wounds, while the use of an anti-HSP90 antibody reversed the effect of ADSC-EXOs; moreover, the neovascularization ability was weaker in diabetic wounds than in non-diabetic wounds (Figure 7E–G). In conclusion, diabetes obstructed wound neovascularization; eHSP90 efficiently promoted neovascularization in diabetic wounds. We performed immunofluorescence staining of Ki67 and immunohistochemical staining of PCNA to evaluate cell proliferation in granulation tissue. The numbers of Ki67-positive cells and PCNA-positive cells were substantially increased after treatment with ADSC-EXOs in diabetic wounds; these effects were not reversed by the use of anti-HSP90 antibody (Figure S9). These results revealed that the mechanism by which ADSC-EXOs promote cell proliferation in diabetic wounds is independent of surface HSP90.
Stem cell transplantation has been investigated in clinical trials as potential treatment for various diseases. Compared with cell therapy, exosome therapy has the following advantages: recapitulation of parental cell biological activity, efficient passage through biological barriers, low immunogenicity, and robust maneuverability [31,32]. There is increasing interest in the use of exosomes for wound-healing therapy, but few studies have focused on the regulatory effects of exosomes on oxidative stress in wounds. Here, we investigated the therapeutic effects of ADSC-EXOs on the modulation of oxidative stress in diabetic wounds along with the underlying mechanisms of such effects. In vitro and in vivo experiments revealed that ADSC-EXOs could protect against hypoxic and oxidative stress injury through surface HSP90, which targeted the LRP1 receptor and activated the AKT signaling pathway, thereby promoting diabetic wound healing (Figure 8). Wound healing is a complex and well-orchestrated biological process; it is commonly divided into four overlapping and sequential phases (hemostasis, inflammation, proliferation, and remodeling) [33]. In the proliferation phase, keratinocytes participate in the construction of new skin epidermis; fibroblasts are an important component cell of formed granulation tissue; and endothelial cells are involved in the formation of new vascular networks [29]. Therefore, the promotion of cell function during this phase can accelerate wound healing. ADSCs constitute a subset of MSCs that reportedly release exosomes, which play important roles in wound repair by regulating inflammation, accelerating cell proliferation and migration, and promoting angiogenesis [34,35]. Here, we successfully extracted high-purity ADSC-EXOs by differential ultracentrifugation; we systematically evaluated the effects of ADSC-EXOs on three key cell types involved in wound healing. Consistent with the findings in previous reports, we showed that ADSC-EXOs could enhance proliferation and migration in HaCaT cells, fibroblasts, and HUVECs in vitro. Moreover, we found that ADSC-EXOs significantly increased the number of capillary-like tubular structures, implying that ADSC-EXOs could enhance angiogenesis. These findings suggest that ADSC-EXOs can promote the functions of three skin cells during wound healing. Oxidative stress is involved in the occurrence and development of diabetic wound healing [5]. In general, the balance of ROS generation and elimination is precisely maintained by intracellular antioxidant protective systems. However, in a state of oxidative stress, excessive ROS leads to redox imbalance in the diabetic wound microenvironment; thus, the endogenous antioxidant system cannot completely prevent damage [36]. The use of exogenous antioxidants has become an effective method to eliminate oxidative stress damage [37]. Antioxidant enzymes (e.g., superoxide dismutase, catalase, glutathione peroxidase, and heme oxygenase), vitamins, medicinal plants, and nanoparticles are used as major antioxidants for wound repair [38]. Recent studies have shown that exosomes can regulate oxidative stress, but the underlying mechanism has been unclear. Li et al. demonstrated that ADSC-EXOs overexpressing Nrf2 can reduce ROS levels and inflammatory cytokine expression levels in endothelial cells exposed to high glucose, thereby promoting endothelial cell survival [39]. Human umbilical cord MSC-derived exosomes have been shown to reverse H2O2 and CCl4-induced oxidative stress damage and apoptosis in liver cells through the delivery of glutathione peroxidase 1 [40]. In the present study, we noted that ADSC-EXOs had a significant antioxidant effect in H2O2-stimulated cells. The ROS levels and cell apoptosis were increased in HaCaT cells, fibroblasts, and HUVECs after H2O2- or COCl2-induced injury; these effects were reversed by ADSC-EXO treatment. Mitochondrial dysfunction is a common feature of delayed wound healing, which causes excessive ROS generation [41]. Previous studies have demonstrated that ADSC-EXOs have beneficial effects on treating liver ischemia reperfusion injury and amyotrophic lateral sclerosis by regulating mitochondrial functions [42,43]. Whether ADSC-EXOs could reduce the intracellular ROS level of skin cells through rescuing mitochondrial functions needs be further studied. Exosomes are cell–cell communication messengers that deliver their contents to recipient cells through three main pathways: ligand–receptor interaction, internalization, and direct membrane fusion [44]. Most exosome research focuses on internalization and direct membrane fusion. For example, exosomes from bone marrow MSCs were internalized by HUVECs [12]. In another study, exosomes from human induced pluripotent MSCs were taken up by HaCaT cells and human dermal fibroblasts [45]. However, in the present study, we detected minimal uptake of PKH26-labeled exosomes by fibroblasts despite incubation for 24 h, while those exosomes were rapidly internalized by 293T cells, HaCaT cells, and HUVECs. Subsequent co-culture experiments involving fibroblasts and Schwann cells confirmed our findings. In addition, we noted that HUVECs formed capillary tubes within 2 h of ADSC-EXO treatment; the AKT and ERK signaling pathways were activated with 30 min of ADSC-EXO treatment. Considering the rapid signaling response, we inferred that ADSC-EXOs might exert effects on wound cells partially through ligand-receptor interactions. Wound healing is a complicated biological process, which requires the accurate cooperation of many different biological and molecular events. Interestingly, the current data show numerous mechanisms are involved in the action of stem cell exosomes on wound healing, including the Notch, ERK, AKT, STAT-3, and Wnt/β-catenin signaling pathways [46]. These pathways may work together and exert multiple functions in different stages of wound healing. Therefore, the effect of ADSC-EXOs on other signaling pathways should be fully investigated in further studies. Ligand–receptor interactions have regulatory roles that mainly involve exosomal membrane proteins, which bind to the recipient cell surface and initiate intracellular signaling pathways [47,48]. To determine which exosomal membrane proteins mediated antioxidant effects, we analyzed proteins present in ADSC-EXOs. Among the 10 most abundant proteins on the plasma membrane, HSP90 was selected for further analysis because it is involved in the wound-healing process. Previous studies have demonstrated that HSP90 is present on the surface of tumor-secreted exosomes and that anti-HSP90 antibody nullified the pro-motility activity of those exosomes [49]. Our results showed that an anti-HSP90 antibody could neutralize the effects of ADSC-EXOs in terms of promoting cell migration and protecting cells against hypoxic and oxidative stress damage. Therefore, we confirmed that ADSC-EXOs mediated their effects on fibroblasts through ligand-receptor interactions, using surface HSP90 as the effector molecule. HSP90 has been reported to bind to LRP1, human epidermal growth factor receptor 2, and TLR4 [24]. LRP1 is a ubiquitous cell-surface receptor that mediates hypoxia-induced cell migration, cancer cell survival, and epithelial-to-mesenchymal transition [24,50]. HSP90 binding to LRP1 has been shown to promote cell migration and survival during wound healing [21,50]. Here, we found that the positive effects of ADSC-EXOs in terms of promoting cell migration and survival were abrogated in LRP1 knockdown fibroblasts, confirming that LRP1 is the membrane receptor for ADSC-EXOs. Extracellular HSP90 binding to LRP1 has been reported to activate multiple signaling pathways, including the AKT, ERK, and NF-κB signaling pathways [49,51]. We found that ADSC-EXOs could activate the ERK and AKT signaling pathways (but not the NF-κB signaling pathway) in fibroblasts; this activation could be inhibited by an anti-HSP90 antibody or by knockdown of LRP1. Furthermore, LY294002 effectively antagonized the effects of ADSC-EXOs in terms of promoting fibroblast migration and protecting fibroblasts against hypoxic and oxidative stress damage, whereas U0126 could not antagonize those effects. Collectively, these results showed that HSP90 on the exosomal surface activated the AKT signaling pathway by binding to the LRP1 receptor; it thus promoted cell migration and improved cell survival under oxidative stress. Finally, we observed the effects of ADSC-EXOs on wound healing in established diabetic wound models. As expected, ADSC-EXOs significantly promoted collagen deposition and neovascularization, thereby facilitating diabetic wound closure; the addition of an anti-HSP90 antibody reversed the effects of ADSC-EXOs. The microenvironment of diabetic wounds is in a state of oxidative stress with high levels of ROS. Previous analyses of stem cells and exosomes have not extensively investigated ROS levels in wounds. We found that ADSC-EXOs exhibited excellent ROS-scavenging and anti-apoptosis performance although these effects were antagonized by the use of anti-HSP90 antibody. However, the anti-HSP90 antibody could not antagonize the effects of ADSC-EXOs in terms of promoting cell proliferation in diabetic wound models.
In this study, we demonstrated that ADSC-EXOs could promote skin cell proliferation, migration, and survival during exposure to hypoxic and oxidative stress damage; they could also facilitate collagen deposition and neovascularization, which led to accelerated diabetic wound healing. The underlying mechanism involved their surface eHSP90, which bound to the LRP1 receptor and activated the downstream AKT signaling pathway. Thus, our findings suggest that the application of ADSC-EXOs can enhance skin healing in diabetic wounds. | true | true | true |
|
PMC9600068 | Constanze Knebel,Roderich D. Süssmuth,Helen S. Hammer,Albert Braeuning,Philip Marx-Stoelting | New Approach Methods for Hazard Identification: A Case Study with Azole Fungicides Affecting Molecular Targets Associated with the Adverse Outcome Pathway for Cholestasis | 19-10-2022 | hepatotoxicity,azole fungicides,molecular targets,adverse outcome pathway,liver cholestasis | Triazole fungicides such as propiconazole (Pi) or tebuconazole (Te) show hepatotoxicity in vivo, e.g., hypertrophy and vacuolization of liver cells following interaction with nuclear receptors such as PXR (pregnane-X-receptor) and CAR (constitutive androstane receptor). Accordingly, azoles affect gene expression associated with these adverse outcomes in vivo but also in human liver cells in vitro. Additionally, genes indicative of liver cholestasis are affected in vivo and in vitro. We therefore analyzed the capability of Pi and Te to cause cholestasis in an adverse outcome pathway (AOP)-driven approach in hepatic cells of human origin in vitro, considering also previous in vivo studies. Bile salt export pump (BSEP) activity assays confirmed that both azoles are weak inhibitors of BSEP. They alternate the expression of various cholestasis-associated target genes and proteins as well as the mitochondrial membrane function. Published in vivo data, however, demonstrate that neither Pi nor Te cause cholestasis in rodent bioassays. This discrepancy can be explained by the in vivo concentrations of both azoles being well below their EC50 for BSEP inhibition. From a regulatory perspective, this illustrates that toxicogenomics and human in vitro models are valuable tools to detect the potential of a substance to cause a specific type of toxicity. To come to a sound regulatory conclusion on the in vivo relevance of such a finding, results will have to be considered in a broader context also including toxicokinetics in a weight-of-evidence approach. | New Approach Methods for Hazard Identification: A Case Study with Azole Fungicides Affecting Molecular Targets Associated with the Adverse Outcome Pathway for Cholestasis
Triazole fungicides such as propiconazole (Pi) or tebuconazole (Te) show hepatotoxicity in vivo, e.g., hypertrophy and vacuolization of liver cells following interaction with nuclear receptors such as PXR (pregnane-X-receptor) and CAR (constitutive androstane receptor). Accordingly, azoles affect gene expression associated with these adverse outcomes in vivo but also in human liver cells in vitro. Additionally, genes indicative of liver cholestasis are affected in vivo and in vitro. We therefore analyzed the capability of Pi and Te to cause cholestasis in an adverse outcome pathway (AOP)-driven approach in hepatic cells of human origin in vitro, considering also previous in vivo studies. Bile salt export pump (BSEP) activity assays confirmed that both azoles are weak inhibitors of BSEP. They alternate the expression of various cholestasis-associated target genes and proteins as well as the mitochondrial membrane function. Published in vivo data, however, demonstrate that neither Pi nor Te cause cholestasis in rodent bioassays. This discrepancy can be explained by the in vivo concentrations of both azoles being well below their EC50 for BSEP inhibition. From a regulatory perspective, this illustrates that toxicogenomics and human in vitro models are valuable tools to detect the potential of a substance to cause a specific type of toxicity. To come to a sound regulatory conclusion on the in vivo relevance of such a finding, results will have to be considered in a broader context also including toxicokinetics in a weight-of-evidence approach.
Cholestasis is a form of substance-induced liver injury that results from an impairment of bile acid excretion causing accumulation of bile acids in the liver and/or the systemic circulation [1]. There are several potential causes of cholestasis, like obstruction of the bile duct, hepatic inflammation or drug–drug interactions [1]. At the molecular level, inhibition of the bile salt export pump (BSEP) is frequently considered as the molecular initiating event (MIE) of the AOP for liver cholestasis, leading to an increase in cellular bile acids and subsequent toxicity [2,3]. Inhibition of BSEP is a frequent cause of substance-induced cholestasis [4]. Several nuclear receptors are involved in bile acid-dependent signaling, especially the constitutive androstane receptor (CAR), the pregnane-X-receptor (PXR), and the farnesoid-X-receptor (FXR) that is activated upon accumulation of bile acids and regulates a number of genes important for bile acid detoxification [5]. The induction of detoxification can be described as an adaptive response, while other events including the induction of mitochondrial dysfunction and oxidative stress can be described as deteriorative responses [6]. A schematic overview is given in Figure 1a that proposes a modified AOP for substance-induced cholestasis based on Vinken et al., 2015 [2]. Triazoles are frequently used fungicides, which display hepatotoxicity in animal studies in rats or mice [7,8]. Hepatocellular hypertrophy is among the most prominent histopathological findings after repeated-dose administration of triazole fungicides, and probably related to activation of a set of nuclear receptors including PXR and CAR, as recently reviewed by Marx-Stoelting et al., 2020 [9]. Some fungicides of this group have also been shown to cause hepatocellular cholestasis: prothioconazole, for example, was reported to be cholestatic in respective in vivo studies conducted for its approval as an active substance for use in pesticides [10]. Moreover, cholestasis is a frequently described finding in drug-induced liver injury caused by pharmacologically used azoles [11]. The ability of azole fungicides to interact with BSEP, the MIE of cholestasis (Figure 1), has not been investigated so far for a number of triazoles, including tebuconazole (Te) and propiconazole (Pi). Since hepatic gene expression potentially related to cholestasis was altered by azoles such as Pi, Te or cyproconazole in previous studies involving transcriptomics analysis in vivo or in vitro [12,13], we decided to further elucidate the potential of Pi and Te to cause cholestasis by application of new approach methods (NAMs). NAMs are under development in several projects for the identification of different forms of hepatotoxicity [12,13,14]. Such methods are considered to help in identifying human-relevant liver effects while at the same time avoiding animal testing. Central question are how well AOP-based batteries of NAMs are capable of predicting adverse effects in vivo, and how well findings are correlated with human disease. For the endpoint liver steatosis, we were recently able to show that this correlation is quite pronounced [15]. For the endpoint cholestasis, the work of others points in the same direction [3]. The aim of our study was to investigate mechanisms of triazole-mediated cholestasis by use of in vitro methods. In addition, the applicability of NAMs for hazard identification in a regulatory context to detect specific forms of drug- or substance-induced liver injury (DILI/SILI) was considered. Therefore, we selected the two triazoles Pi and Te as test substances in in vitro assays by using the human liver cell lines HepG2 and HepaRG. On top of that we investigated the compounds in a binary mixture to see if the mixture response was in line with the assumption of concentration addition for substances with a similar mode of action. Besides BSEP inhibition, reporter gene assays were used to analyze nuclear receptor transactivation relevant for cholestasis. Additionally, we measured gene and protein expression as well as the level of mitochondrial membrane disruption. Results were finally compared to existing in vivo results for various azole fungicides used as pesticides or drugs, based on test guidelines generally applied in regulatory hazard assessment to allow the applicability of the NAM for regulatory purposes to be assessed. Concentration inducing effects in vitro were also compared to concentrations measured in vivo.
Technical grade Pi (CAS # 60207-90-1; Batch # CGA64250B; purity 96.10%) was purchased from Syngenta (Basel, Switzerland) and technical grade Te (CAS # 107534-96-3; Batch # NK21BX0392; purity 96.20%) from Bayer (Leverkusen, Germany). Dimethylsulfoxide (DMSO) was used as solvent for the test compounds and added to the cells resulting in a final DMSO concentration of 0.2% (v/v). Equimolar mixtures of Pi and Te were applied in mixture trials (i.e., “5 μM Pi+Te” corresponds to 2.5 μM Pi + 2.5 μM Te). Table 1 shows structural formulas of the used compounds.
Undifferentiated HepaRG cells (Biopredic International, Saint Grégoire, France) were seeded and cultured as previously described [16,17]. In brief, cells were cultured in William’s E medium with 2 mM glutamine (Pan-Biotech, Aidenbach, Germany) supplemented with 10% fetal calf serum (FCS; Pan-Biotech), 100 U/mL penicillin and 100 μg/mL streptomycin (Capricorn Scientific, Ebsdorfergrund, Germany), 0.05% human insulin (PAA Laboratories GmbH, Pasching, Austria) and 50 μM hydrocortisone hemisuccinate (Sigma-Aldrich). After cultivation for two weeks, cells were differentiated in the medium mentioned above containing 1.7% DMSO in addition for another two weeks. Differentiated HepaRG cells were treated with test compounds in treatment medium (phenol red-free Williams E medium, Pan-Biotech, supplemented with the same supplements as the differentiation medium, but only 2% FCS and 0.2% DMSO) for 24 h or six days with renewal of medium/treatments every two days. Human HepG2 hepatocellular carcinoma cells (ECACC, Salisbury, UK) were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Pan-Biotech) supplemented with 10% FCS (Pan-Biotech) as described previously [16]. Treatment with test substances was conducted in phenol red-free DMEM medium (Pan-Biotech) supplemented with 10% FCS for 24 h. A Binder cell culture incubator was used for incubation of both cell lines at 37 °C and 5% CO2 in a humidified atmosphere.
Cytotoxicity of Pi and Te was analyzed using the colorimetric MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) reduction assay in HepG2 and HepaRG cells in 96-well format according to standard protocols [18]. The detergent Triton X-100 (0.01%) served as positive control. Concentrations of 40 μM (24 h treatment) and 20 μM (6 days treatment), yielding ≥80% cell viability, were chosen as the maximum concentrations of Pi and Te for further experimentation to ensure the absence of artifacts caused by cytotoxicity.
Gene expression analysis was conducted as previously described [7,19] The human microarray Agilent Expression Profiling Service (incl. 8 × 60K Array) was conducted by ATLAS Biolabs GmbH (Berlin, Germany) as previously described [12,19] using RNA from HepaRG cells treated with a combination of 10 μM Pi and 10 μMTe for 24 h. Results (fold changes of each treatment relative to the solvent control, 0.2% DMSO) were further evaluated using the bioinformatics analysis and search tool IPA (Ingenuity Pathway Analysis) from QIAGEN (Germantown, MD, USA) using the IPA “Tox Analysis” tool. A p-value < 0.05 and a |fold change| > 2 were used as cutoff criteria for the transcriptomics data. In IPA, standard settings (no filtering, direct as well as indirect relationships were considered) were selected (date of analysis: 12 October 2018) [12]. PCR-based gene expression analysis was conducted as recently described [7,19] using the primer pairs listed in Supplementary Table S4.
Dual luciferase reporter gene assays for CAR, PXR and FXR were conducted to analyze the capability of Pi and Te to activate these nuclear receptors in HepG2 cells. The plasmids and assays have been described in detail before [16]. In brief, HepG2 cells were cultivated in 96-well plates and transiently transfected with plasmids (Supplementary Table S5) using TransIT-LT1 (Mirus Bio LLC, Madison, WI, USA) in a relation of 3:1 (TransIT-LT1 [μL]: amount of plasmids [μg]). For the FXR transactivation assay, the first plasmid is based on a fusion protein of GAL4 with the ligand-binding domain (LBD) of FXR. The second plasmid contains a firefly luciferase reporter gene under control of the GAL4-specific upstream activation sequence (UAS). For the CYP7A1 promoter assay, the luciferase reporter construct is driven by a fragment of the promoter of the human FXR-responsive CYP7A1 gene [14]. All measurements were performed according to the Dual Luciferase Assay protocol as provided by the manufacturer (Promega, Madison, WI, USA) and detailed elsewhere [19,20] by using a plate reader (Infinite M200PRO, Tecan, Männedorf, Switzerland).
The assay was conducted by Solvo (Szeged, Hungary) according to standard protocols. In brief, it utilizes isolated cell membrane preparations from HEK293 cells overexpressing human BSEP. The probe substrate was incubated in the presence of ATP (active transport condition) or AMP (passive diffusion condition) in triplicates (n = 3). Control incubations with a known BSEP substrate (taurocholate), as well as taurocholate together with the known BSEP inhibitor cyclosporine A, were conducted alongside the incubations.
Expression levels of transport proteins were investigated using a so-called Triple X Proteomics (TXP) targeted proteomic analysis as described in a previous study [15]. In essence, cell pellets were lysed in buffer for one hour, and the protein concentration was subsequently determined by the bicinchoninic acid (BCA) assay (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s manual. Next, proteolysis was performed overnight with trypsin. Endogenous as well as stable isotope-labeled reference peptides were precipitated using TXP antibodies (customized production by Pineda, Berlin, Germany) using magnetic beads coated with protein G (Thermo Fisher Scientific). The precipitated peptides were then quantified using a previously described 10 min LC-MS (liquid chromatography coupled to mass spectrometry) method with an UltiMate 3000 RSLCnano and a tSIM-QExactive Plus™ mass spectrometer (Thermo Fisher Scientific). Peak areas of the known amounts of the isotope-labeled peptides were set in relation to endogenous signals at the parent ion level [21].
Another key event (KE) in the cholestasis AOP is the alteration of the mitochondrial membrane function. The JC-1-assay was used to address this KE via measurement of the voltage-dependent accumulation of the charged dye JC-1 (5,5′,6,6′tetrachloro- 1,1′,3,3′-tetraethylbenzimidazol-carbocyanine iodide). JC-1 enters the mitochondria and accumulates depending on the membrane potential. For high membrane potential, JC-1 aggregates and emits red light (595 nm). Conversely, it emits green light (535 nm) at low membrane potential. For the measurements, HepaRG cells were cultivated in 96-well plates and treated with the test substances Pi and Te for six days. Valinomycin (10 μM) was applied for 24 h and served as positive control. A volume of 100μL JC-1-solution was added to the cells. After an incubation time of 20 min and washing twice with PBS, fluorescence emission was measured. The ratio of red/green (595 nm/535 nm) fluorescence values were calculated and finally normalized to the solvent control.
In vivo results from previous studies of our group [7,8] as well as from regulatory guideline studies as summarized by Nielsen et al. in 2012 [10] were considered to evaluate whether cholestasis could be observed at the histopathological level in rodents following repeated-dose treatment with Pi and Te. Human in vivo drug side effects by azoles were obtained by searching the European drug vigilance database for the substances fluconazole, itraconazole and ketoconazole, at http://www.adrreports.eu/en/index.html (accessed on 8 June 2021).
Statistical analysis was conducted with SigmaPlot for Windows software (Version 14.0). Shapiro–Wilks and Brown–Forsythe tests were respectively used to analyze if results were normally distributed and for homogeneity of variances. Since most of the data did not meet the prerequisites for parametric testing, the non-parametric Mann–Whitney rank sum test was applied. An asterisk (*) indicates statistical significance at p < 0.05, and error bars depict the standard deviation. IC50 values were calculated by linear regression from measured values with SigmaPlot for Windows software.
In the course of an evaluation of microarray gene expression data from a previous study on the combined effects of Pi and Te in HepaRG human hepatocarcinoma cells, bioinformatic analyses indicated that the compounds exerted transcriptional changes potentially related to hepatic adverse outcomes (Figure 1b; see also Knebel et al., 2019 [12]). This approach was chosen as HepaRG cells constitute an in vitro system closely resembling human hepatocytes (see also discussion section for more information), and since our own previous analyses (e.g., see [12]) have proven the usefulness of bioinformatic analyses of omics data as a NAM to obtain information about liver toxicity and its mode of action. Of note, the majority of toxicity-relevant functions affected by the compounds, as determined by bioinformatic analysis, showed a close relationship to the activation of nuclear receptors. For example, predicted effects of Pi and Te such as liver carcinoma formation, hepatic steatosis, hyperplasia, and liver enlargement are prototypical consequences of CAR and PXR activation in liver cells. Interestingly, transcriptomic data also predicted hepatic cholestasis as an outcome of Pi and Te treatment of HepaRG cells. Liver cholestasis is an adverse outcome also related to nuclear receptor activation, including not only the activation of PXR and CAR (Figure 1a), but also the farnesoid-X-receptor (FXR). The present study therefore aimed to elucidate the potential of both compounds to induce cholestatic effects in human liver cells, as well as to perform comparative analyses in order to assess the in vivo relevance of the findings. The results are presented in accordance with the AOP proposed in Figure 1a. Therefore, NAMs were used to first check the MIE (BSEP inhibition), followed by reporter gene assays for the nuclear receptors involved. Subsequently, the expression of genes and proteins as well as the mitochondrial membrane function for the respective KE were checked. Finally, the results obtained by NAMs in vitro were compared to in vivo results from previous studies, including compound concentrations measured in vivo as reported in a previous publication [8].
Activity of BSEP was analyzed using an in vitro approach based on membrane preparations of human HEK293 cells as a NAM tailored specifically to detect interaction with this particular transport protein. We found that BSEP was inhibited by both compounds: Pi inhibited BSEP-mediated taurocholate accumulation in a concentration-dependent manner with a maximum inhibition of 91% at a concentration of 300 μM (Figure 2a). The calculated IC50 was 78.56 μM (Table 2). Te inhibited BSEP-mediated taurocholate accumulation in a concentration-dependent manner with a maximum inhibition of 99% at a concentration of 300 μM (Figure 2a). The calculated IC50 was 38 μM (Table 2). Both compounds should therefore be regarded weak to moderate inhibitors of BSEP, which generally is in line with the assumption that inhibition of BSEP is the MIE of the AOP for hepatic cholestasis. In addition, expression of the ABCB11 gene (encoding BSEP) was significantly down-regulated at the mRNA level in a concentration-dependent manner in HepaRG cells by Pi alone (significant at the highest dose level only) and by the equimolar mixture of Pi and Te (Figure 2b).
According to the AOP, BSEP inhibition will lead to subsequent activation of the nuclear receptors PXR, CAR, and FXR (Figure 1a). This was analyzed using reporter gene assays in HepG2 cells, due to limitations of HepaRG cells with regard to transfectability. The reporter systems used as NAMs here measure the potential of fusion proteins of the respective receptor and the GAL4 protein, brought into the cells via transient transfection of plasmid DNA (e.g., see [12]). Activating binding of the test compound to the nuclear receptor will induce transcription of a luciferase reporter gene. Both compounds activated PXR (Table 1; Supplementary Table S3; see also methodology and data in Refs. [12,19]). On the other hand, Pi and Te exerted opposite effects on CAR, with Pi acting as an activator and Te as an inhibitor (Supplementary Table S3; see also methodology and data in Refs. [12,19]). Please note that data on PXR activation have been published previously [12]. With respect to FXR, statistically significant receptor activation was observed after Pi treatment, but not with Te (Figure 3a). Thus, in summary, PXR was consistently activated by both test compounds, whereas the effects on CAR (Pi inducer, Te inhibitor) and FXR (Pi inducer, Te no effect) were inconsistent.
In line with assumptions from the AOP, Pi and Te provoked several alterations in the expression of genes associated with cholestasis in HepaRG cells. Alterations observed at the mRNA level are summarized in Figure 4b. For details please refer to Supplementary Table S1. Among these was a down-regulation of CYP7A1, the rate-limiting key enzyme in bile acid synthesis (for details see Figure 3c). This was confirmed using a luciferase reporter system driven by a fragment of the human CYP7A1 promoter in HepG2 cells (Figure 3b). Additional transcriptional changes were observed e.g., for CYP8B1, which was down-regulated by Pi and Te, as well as for the up-regulated genes ABCC3 and ABCG5 (Figure 4a). CYP8B1 is an enzyme involved in bile acid synthesis, while the transporters ABCC3 and ABCG5 are involved in the cellular export of bile acids and cholesterol. While most of these effects were observed as an early response to exposure 24 h after the start of incubation, some were also observable after 6 days of treatment. Equimolar mixtures of both azoles led to transcriptional effects similar to the individual compounds. Figure 4b gives an overview of all genes affected that are responsible for bile acid excretion or synthesis. For some of the proteins encoded by the aforementioned genes, mass-spectrometric assays for protein determination were available. Thus, protein level confirmation of some of the observed changes was achieved (Figure 4b). Detailed protein data are presented in Supplementary Table S2.
To assess possible disruption of the mitochondrial membrane, a further KE in the cholestasis AOP, the JC-1-assay was performed. This NAM is based on a fluorescent dye which accumulates in mitochondria depending on their membrane potential. We found that Pi and Te alone as well as their equimolar combination depolarize the mitochondrial membrane (Figure 5). Mitochondrial depolarization was presumably due to a disruption of the cellular organelle causing the mitochondrial permeability transition pores to open. In turn, the equilibrium of charges was gradually achieved and thus the membrane potential decreased. This decrease is in line with the assumption made in the AOP of cholestasis that mitochondrial function is altered.
In order to allow for a comparison of in vitro and in vivo cholestatic effects of triazoles, published data were reviewed for cholestasis-relevant observations with triazoles applied in agriculture or as pharmaceuticals. For the pesticidal active substances Pi and Te, in vivo effects observed in rodent bioassays after short- and long-term oral exposure were hepatocellular hypertrophy and vacuolization, as well as slight increases in liver enzymes in clinical chemistry (for details see [9]). However, cholestasis was not observed in vivo in guideline-compliant repeated-dose toxicity studies for either of the two substances (see Table 1). For other active substances from the azole group, a summary of in vivo studies was checked for toxicological findings by Nielsen et al., 2012 [10]: three active substances were identified that caused cholestasis in vivo, namely, difenoconazole, prothioconazole and triticonazole. However, the proportion of substances from the azole group causing cholestasis appears rather small, as there are at present more than 20 azole compounds used as active substances in agriculture.
For drugs, the European database on pharmacovigilance was checked for reports on the clinically used azole fungicides itraconazole, ketoconazole and fluconazole (www.adrreports.eu; accessed on 8 June 2021). For itraconazole, approximately 900 adverse drug reactions (ADRs) were reported affecting the hepatobiliary system in the ADR database until 8 June 2021. Of these, only 22 cases showed a cholestatic phenotype (cholestasis or cholestatic liver injury). Jaundice was observed in 56 cases, jaundice cholestatic in only 5 cases. Hyperbilirubinaemia was observed in 16 cases. For ketoconazole, approximately 350 adverse drug reactions were reported affecting the hepatobiliary system in the ADR database until 8th June 2021. Of these only 12 cases showed a cholestatic phenotype (cholestasis or cholestatic liver injury). Jaundice was observed in 37 cases. Hyperbilirubinaemia was observed in 3 cases. For fluconazole, approximately 1100 adverse drug reactions were reported affecting the hepatobiliary system in the ADR database until 8 June 2021. Of these approximately 110 cases showed a cholestatic phenotype (cholestasis or cholestatic liver injury). Jaundice was observed in 84 cases, jaundice cholestatic in 14 cases. Hyperbilirubinaemia was observed in 32 cases. Overall, these results indicate that some azoles generally have the potential to cause cholestasis in vivo in rodent bioassays or in humans when used as drugs for treatment of mycosis. However, in comparison to other hepatobiliary effects of azoles, cholestasis appears to occur at a relatively low frequency.
In vitro transcriptomics analysis revealed cholestasis as one of the top pathways affected after treatment of human liver-derived HepaRG cells with Pi and Te. This is in line with transcriptomics findings for other azoles like cyproconazole, epoxiconazole or prochloraz in vitro and in vivo [7,13,22]. Considering the molecular AOP concept, the present observations with Pi and Te go well in line with the assumption of a cholestatic potential of azoles, as the MIE (inhibition of the activity of BSEP) was clearly affected, and also because effects on nuclear receptors were recorded, especially a consistent and robust activation of PXR by both test compounds. In addition, diminished expression of ABCB11—the gene encoding BSEP—was observed at the mRNA level, adding an additional aspect to the mode of action of azoles to interfere with the AOP for liver cholestasis. This is an interesting observation: the literature is not consistent regarding the effects of PXR, FXR and CAR ligands towards BSEP expression, reporting either induction or no substantial effects [23,24]. Therefore, the lowered levels of BSEP mRNA might point towards additional molecular mechanisms of Pi and Te beyond nuclear receptor activation. However, no cholestasis was observed for Pi and Te in a number of in vivo studies summarized by Nielsen et al., 2012 [10]. This was also confirmed by the absence of such an adverse effect in a number of regulatory studies performed according to OECD test guidelines within the approval processes of the active substances in Europe (summarized in the respective EFSA conclusions [25,26]). Is this an indication that the NAM-based testing or the AOP concept may fail? Several explanations related to the sensitivity of the used cell lines as compared to primary hepatocytes or the in vivo situation could be discussed in this context to explain the observed discrepancy, as could incomplete correlation between mRNA analyses and actual protein levels, as well as sensitivity of the applied methods. However, the HepaRG model has been compared to primary hepatocytes and in vivo models, and found to be of comparable sensitivity and functionality with respect to many liver-specific features [27,28]. Therefore, even if the above reasons cannot fully be ruled out as underlying causes, a different explanation seems much more likely. Indeed, on the one hand, neither of the two azoles under investigation caused cholestasis in standard rodent bioassays in vivo. Considering this, the transcriptomics approach is suspect of being over-predictive. On the other hand, one cannot deny that substances of the azole fungicide group bear a certain potential to cause cholestasis, as shown for some other agricultural fungicides of the same class in vivo [10] and as also documented by cases of DILI, pointing towards a cholestatic potential of some azole fungicides used as human drugs (as documented in the ADR database and summarized in Section 3.5 above). The present results indicate that Pi and Te are both capable of inducing molecular changes in human liver cells, which are indicative of cholestasis, and which are in line with the assumptions of the respective AOP (e.g., BSEP inhibition, PXR activation). In order to judge the in vitro findings, it is crucial to consider different aspects of toxicodynamics and toxicokinetics, which might explain the apparent gap between the different findings mentioned above. These may be species differences, difficulties of in vitro–in vivo extrapolation or kinetic differences. In general, species differences may be the underlying cause of a divergence between rodent in vivo data and results from in vitro experimentation with human cells [29,30]. For several azole fungicides, for example, a non-genotoxic mechanism of liver tumor induction via activation of the receptor CAR has been identified, the relevance of which for humans is controversial [29,30]. However, the fact that some azoles can act in a cholestatic manner in laboratory rodents and certain ones also in humans indicates that the chemical class of azoles generally has the capability of exerting cholestatic effects in both species, making a scenario of fundamental species differences for this particular endpoint appear rather unlikely. A general in vitro–in vivo discrepancy for the ability of azoles to exert cholestatic effects in humans or rodents does also appear very unlikely, as the observation that azoles may cause DILI in human patients treated for mycosis is generally in line with the in vitro observations in cultured human liver cells. This has been observed in rare cases for itraconazole, ketoconazole or fluconazole as reported in the EU ADR database analyzed in Section 3.5. Moreover, a small proportion of the azole fungicides used in agriculture (e.g.,difenoconazole or prothioconazole) have shown a cholestatic potential in vivo [10]. In addition, the cell lines used have successfully been used by others on several cholestatic substances [3,31]. Instead, the main reason for the non-occurrence of cholestasis with Pi and Te in vivo seems to be kinetic in nature, and related to the cellular levels of the azole compounds achieved in the different types of experiments: as shown in Table 1, the IC50 for BSEP inhibition is ~5–50 fold above the intra-hepatic concentration measured at the top dose in previous animal studies [8]. Thus, assuming that the inhibition of murine BSEP by Pi and Te occurs at similar concentrations as for human BSEP, the tissue levels needed to induce BSEP inhibition in vivo have probably not been reached in the respective studies. Nuclear receptor activation, as exemplified by PXR activation and associated with other liver effects such as CYP induction, hypertrophy and hepatocellular proliferation, occurs in a dose range similar to the observed in vivo concentrations (see Table 1), explaining why the latter forms of nuclear receptor-mediated liver effects are observed in vivo. In line with the AOP for hepatic cholestasis, PXR activation alone appears not to be sufficient to trigger the adverse outcome, whereas BSEP inhibition is most likely needed. Our results illustrate both strengths and weaknesses of in vitro-based NAM assays involving toxicogenomics. On the positive side, these methods accurately detect certain hazards and allow for screening of potential effects without the need for animal testing. On the other side, NAMs may be over-predictive, especially when aspects of kinetics are not considered. It appears most likely that the potential of Pi and Te to induce cholestasis is, in general, correctly predicted by NAM in vitro indicating the usefulness of these NAMs. However, parameters such as in vivo tissue concentrations and compound potencies have to be considered carefully when using these data to predict adverse outcomes in vivo (as demonstrated in Table 1). Hence, if not used for definitive prediction of an adverse outcome on its own but either in combination with other approaches or for prioritization for further testing, the approach is well suited to fit into a tiered regulatory procedure as an early step to predict a hazard. In a later phase of such an approach, specific organ concentrations, known from in vivo studies or computed using appropriate toxicokinetic models including in silico models, need to be considered in order to conclude on the risk of an adverse outcome [32]. A number of in silico physiologically-based pharmacokinetic (PBPK) models exist that may in principle be used, as recently reviewed by Ref. [33]. In the present study, this was not the case as the concentrations needed for in vitro inhibition of BSEP exceed the tissue concentrations achievable in rodent in vivo studies. In this context it should be mentioned that the concentrations triggering cholestasis in vitro are also several orders of magnitude above realistic human exposure levels, as maximum residue levels in food are well below the effect dose levels.
The present work demonstrates the ability of NAM-based in vitro testing involving toxicogenomics to correctly identify hazardous properties of substances. It also shows the necessity to combine such methods with considerations of toxicokinetics to come to sound conclusions in risk assessment. | true | true | true |
|
PMC9600248 | 36125275 | Long C. Nguyen,David M. Renner,Diane Silva,Dongbo Yang,Nicholas A. Parenti,Kaeri M. Medina,Vlad Nicolaescu,Haley Gula,Nir Drayman,Andrea Valdespino,Adil Mohamed,Christopher Dann,Kristin Wannemo,Lydia Robinson-Mailman,Alan Gonzalez,Letícia Stock,Mengrui Cao,Zeyu Qiao,Raymond E. Moellering,Savas Tay,Glenn Randall,Michael F. Beers,Marsha Rich Rosner,Scott A. Oakes,Susan R. Weiss | SARS-CoV-2 Diverges from Other Betacoronaviruses in Only Partially Activating the IRE1α/XBP1 Endoplasmic Reticulum Stress Pathway in Human Lung-Derived Cells | 20-09-2022 | IRE1α pathway,MERS-CoV,OC43,SARS-CoV-2,coronavirus,unfolded protein response | ABSTRACT Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has killed over 6 million individuals worldwide and continues to spread in countries where vaccines are not yet widely available or its citizens are hesitant to become vaccinated. Therefore, it is critical to unravel the molecular mechanisms that allow SARS-CoV-2 and other coronaviruses to infect and overtake the host machinery of human cells. Coronavirus replication triggers endoplasmic reticulum (ER) stress and activation of the unfolded protein response (UPR), a key host cell pathway widely believed to be essential for viral replication. We examined the master UPR sensor IRE1α kinase/RNase and its downstream transcription factor effector XBP1s, which is processed through an IRE1α-mediated mRNA splicing event, in human lung-derived cells infected with betacoronaviruses. We found that human respiratory coronavirus OC43 (HCoV-OC43), Middle East respiratory syndrome coronavirus (MERS-CoV), and murine coronavirus (MHV) all induce ER stress and strongly trigger the kinase and RNase activities of IRE1α as well as XBP1 splicing. In contrast, SARS-CoV-2 only partially activates IRE1α through autophosphorylation, but its RNase activity fails to splice XBP1. Moreover, while IRE1α was dispensable for replication in human cells for all coronaviruses tested, it was required for maximal expression of genes associated with several key cellular functions, including the interferon signaling pathway, during SARS-CoV-2 infection. Our data suggest that SARS-CoV-2 actively inhibits the RNase of autophosphorylated IRE1α, perhaps as a strategy to eliminate detection by the host immune system. | SARS-CoV-2 Diverges from Other Betacoronaviruses in Only Partially Activating the IRE1α/XBP1 Endoplasmic Reticulum Stress Pathway in Human Lung-Derived Cells
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has killed over 6 million individuals worldwide and continues to spread in countries where vaccines are not yet widely available or its citizens are hesitant to become vaccinated. Therefore, it is critical to unravel the molecular mechanisms that allow SARS-CoV-2 and other coronaviruses to infect and overtake the host machinery of human cells. Coronavirus replication triggers endoplasmic reticulum (ER) stress and activation of the unfolded protein response (UPR), a key host cell pathway widely believed to be essential for viral replication. We examined the master UPR sensor IRE1α kinase/RNase and its downstream transcription factor effector XBP1s, which is processed through an IRE1α-mediated mRNA splicing event, in human lung-derived cells infected with betacoronaviruses. We found that human respiratory coronavirus OC43 (HCoV-OC43), Middle East respiratory syndrome coronavirus (MERS-CoV), and murine coronavirus (MHV) all induce ER stress and strongly trigger the kinase and RNase activities of IRE1α as well as XBP1 splicing. In contrast, SARS-CoV-2 only partially activates IRE1α through autophosphorylation, but its RNase activity fails to splice XBP1. Moreover, while IRE1α was dispensable for replication in human cells for all coronaviruses tested, it was required for maximal expression of genes associated with several key cellular functions, including the interferon signaling pathway, during SARS-CoV-2 infection. Our data suggest that SARS-CoV-2 actively inhibits the RNase of autophosphorylated IRE1α, perhaps as a strategy to eliminate detection by the host immune system.
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) emerged in China in late 2019. It was the third lethal zoonotic coronavirus to emerge into humans, after SARS-CoV (2002) and Middle East respiratory syndrome coronavirus (MERS-CoV) (2012), each of which has been associated with acute lung injury and hypoxemic respiratory failure. While coronaviruses are divided into four genera (alpha, beta, gamma, and delta) (1, 2), all three of the lethal human coronaviruses are betacoronaviruses, albeit from different subgenera (Fig. 1). SARS-CoV and SARS-CoV-2 are sarbecoviruses, while MERS-CoV is a merbecovirus. Other human CoVs, including HCoV-OC43 (OC43) and HCoV-HKU1 (HKU-1), are embecoviruses, as is the model murine coronavirus mouse hepatitis virus (MHV). All CoVs have similar genome structures and replication cycles, and the human CoVs as well as some MHV strains exhibit tropism for the epithelia of the respiratory tract, the portal of entry. They replicate their RNAs and produce subgenomic mRNAs by conserved mechanisms and encode homologous structural as well as replicase proteins. Despite the similarities among all coronaviruses, each subgenus expresses distinct accessory proteins that may confer differences in host-virus interactions. Indeed, we have previously found that SARS-CoV-2, MERS-CoV, and MHV all induce somewhat different levels of activation and/or antagonism of interferon (IFN) signaling and other double-stranded RNA (dsRNA) induced antiviral innate responses (3–11). One key pathway involved in the virus-induced host response is the endoplasmic reticulum (ER) stress response that regulates protein homeostasis (referred to as proteostasis) in this organelle. One-third of all eukaryotic proteins, including most that are inserted into membranes or secreted, are synthesized through co-translational translocation into the ER lumen. Likewise, viral membrane-associated proteins are translated and processed in association with the ER (12, 13). Once in the ER, these polypeptides undergo stringent quality control monitoring to ensure that they are properly processed and folded. If the capacity to fold proteins is unable to keep up with demand, misfolded proteins will accumulate in the ER lumen—a condition referred to as “ER stress.” The presence of misfolded proteins in the ER is sensed by three transmembrane sentinel proteins—activating transcription factor 6 (ATF6), PKR-like ER kinase (PERK), and inositol-requiring enzyme (IRE)1α—which trigger an intracellular signaling pathway called the unfolded protein response (UPR). In an effort to restore proteostasis, activation of these sensors induces transcription factors that turn on genes encoding chaperones, oxidoreductases, and ER-associated decay (ERAD) components (14). The UPR also inhibits cap-dependent translation, thus decreasing the load on the ER and giving it extra time to fold proteins already in production (15, 16). If successful, these adaptive UPR programs restore ER homeostasis. The most ancient UPR pathway is controlled by IRE1α—an ER transmembrane bifunctional kinase/endoribonuclease (RNase) that employs autophosphorylation to control its catalytic RNase function (17, 18). In response to ER stress, IRE1α undergoes autophosphorylation and dimerization to allosterically activate its RNase domain to excise a 26-nucleotide (nt) nonconventional intron in XBP1 mRNA; religation of spliced XBP1 shifts the open reading frame, and its translation produces the homeostatic transcription factor XBP1s (s = spliced) (19, 20). Once synthesized, XBP1s upregulates genes that expand the ER and its protein folding machinery (21). IRE1α can additionally lead to apoptosis and inflammation via JUN N-terminal kinase (JNK) and p38 mitogen-activated protein kinase (MAPK) signaling (22). Prolonged ER stress can induce regulated IRE1-dependent decay (RIDD), promoting the cleavage of additional targets beyond XBP1 mRNA, such as secretory protein and ER-localized mRNAs (23). In the short term, RIDD may promote adaptation through further reducing translation and the protein burden on the ER. However, prolonged RIDD leads to the depletion of vital ER resident enzymes and structural components to exacerbate ER stress and hasten cell death (17, 24). There is a large body of evidence that viral replication in mammalian cells can trigger ER stress and UPR activation in infected cells (25), and numerous studies report that the UPR is activated upon infection of host cells by coronavirus family members (12, 13, 26–31). Coronaviruses induce stress in the ER in several ways. First, conserved replicase-encoded, nonstructural proteins nsp3, nsp4, and nps6 are embedded into the ER membrane and, along with unknown host factors, promote membrane curvature to form double membrane vesicles (DMVs), the site of viral replication/transcription centers (RTC) (32). In addition to remodeling the ER, coronaviruses further condition infected cells by shifting translation away from host mRNAs and instead to viral mRNAs. Translation of viral mRNAs causes the ER to be flooded with heavily glycosylated viral structural proteins (e.g., spike [S], membrane [M], and envelope [E]), challenging the organelle’s folding capacity and overall integrity. Indeed, overexpression of coronavirus spike proteins (33) as well as several sarbecovirus accessory proteins (28, 34), has been reported to induce ER stress, although overexpression itself may cause stress irrespective of the proteins. Finally, cell membranes are depleted as enveloped virus particles are assembled into new virions in the ER-Golgi intermediate compartment before budding from the infected cell (1). Thus, coronaviruses as well as other enveloped viruses promote a massive ER expansion and modification necessary to replicate their genomes, transcribe mRNAs, and finally, to process and package their protein products into viral particles. We have compared the activation status and requirement of the IRE1α/XBP1 arm of the UPR in well-characterized human lung epithelial cell lines and in induced pluripotent stem cell (iPSC)-derived type II alveolar (iAT2) cells, following infection with four betacoronaviruses representing three distinct subgenera. We find that infection with MERS-CoV, OC43, and MHV leads to phosphorylation of IRE1α and the consequent production of spliced XBP1 (XBP1s) transcription factor. Surprisingly, while we observed phosphorylation of IRE1α in SARS-CoV-2 infected cells, there was a notable absence of XBP1s, suggesting that SARS-CoV-2 inhibits downstream signaling of the IRE1α/XBP1 arm of the UPR. In addition, we report reduced SARS-CoV-2-induced interferon signaling gene expression in the absence of IRE1α.
To determine whether betacoronaviruses activate IRE1α, we first examined the level of phosphorylated IRE1α after viral infection of the A549 human lung carcinoma cell line. We used A549 cells stably expressing the following receptors to facilitate optimal entry for each of the viruses: carcinoembryonic antigen cell adhesion molecule (CEACAM) 1a or MHVR (MHV), dipeptidyl peptidase DPP4 (MERS-CoV) or angiotensin-converting enzyme 2 (ACE2) (SARS-CoV-2). HCoV-OC43 can infect parental A549 or cells expressing ACE2 (3). Consistent with previous reports that embeco subgenus coronaviruses MHV (26, 35) and OC43 (30) induce ER stress, we observed a significant increase in phospho-IRE1α (p-IRE1α) during infection by either OC43 (24 or 48 h postinfection [hpi]) or MHV (24 hpi) (Fig. 2A to C). To confirm the specificity of the p-IRE1α band, we pretreated cells prior to infection with KIRA8, a highly selective kinase inhibitor of IRE1α known to inhibit both autophosphorylation and, consequently, RNase activity. As expected, KIRA8 significantly inhibited the induction of p-IRE1α by OC43 and MHV (Fig. 2A and C). Thapsigargin (Tg) and tunicamycin (TM), both inducers of ER stress, were used as further controls (Fig. 2B and D, and E). Robust induction of p-IRE1α was observed with 1 h of Tg (1 μM) treatment, while no activation of p-IRE1α was observed after 8 h of treatment with TM (1 μg/mL), consistent with the negative feedback regulation observed with extended TM treatment (36). We also observed robust phosphorylation of IRE1α in A549-DDP4 cells and A549-ACE2 cells infected by MERS-CoV and SARS-CoV-2, respectively, at 24 and 48 hpi (Fig. 2D to F and Fig. S1A and B in the supplemental material). As with OC43 and MHV, IRE1α phosphorylation during SARS-CoV-2 infection was inhibited by KIRA8 (Fig. 2F). Interestingly, we observed a decrease in OC43 (Fig. 2A) and SARS-CoV-2 (Fig. 2F) nucleocapsid expression in KIRA8-treated cells. However, this may be the result of off-target effects from the compound rather than solely from IRE1α inhibition, given our findings described below using IRE1α knockout (KO) cells. These results are not limited to a single cell type, as we observed similar induction of p-IRE1α in Calu-3 cells, another lung epithelial-derived cells line, which can be productively infected with both MERS-CoV or SARS-CoV-2 (Fig. 2G). These results demonstrate that MERS-CoV, SARS-CoV-2, HCoV-OC43, and MHV activate the host IRE1α kinase after infection. 10.1128/mbio.02415-22.1 Kinetics of activation of IRE1α phosphorylation during infection with MERS-CoV or SARS-CoV-2. (A and B) A549 cells expressing the indicated viral receptors were mock infected or infected with MERS-CoV (A) or SARS-CoV-2 (B) at an MOI of 5. At the indicated time points, total protein was harvested and analyzed by immunoblotting with the indicated antibodies. Cells treated with thapsigargin (Tg; 1μM) for 1 h or tunicamycin (TM; 1 μg/mL) for 8 h were used as a positive control for IRE1α phosphorylation and attenuation, respectively. Data shown are from one representative experiment from at least two independent experiments. Download FIG S1, PDF file, 0.3 MB. Copyright © 2022 Nguyen et al. 2022 Nguyen et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
We next examined the effect of coronavirus infection on the RNase activity of IRE1α as assessed by XBP1 splicing. Using specific primers to quantify spliced XBP1 mRNA (XBP1s), we observed a marked increase in the percentage of spliced XBP1 mRNA (%XBP1s) as well as an increase in the relative amount of spliced XBP1 mRNA (XBP1s) compared to the mock control after infection by OC43, MERS-CoV, or MHV in receptor-expressing A549 cells (Fig. 3A and B and Fig. S2A and B). This induction of XBP1s by OC43 and by MERS-CoV infection was confirmed by assessing XBP1 splicing by agarose gel electrophoresis (Fig. 3E and F). DNAJB9, a canonical target of XBP1s, was also markedly upregulated with OC43, MERS-CoV, and MHV infection at both 24 and 48 hpi (Fig. 3A and B and Fig. S2B). This induction of IRE1α RNase activity is coincident with the observed autophosphorylation of p-IRE1α upon OC43, MHV, or MERS-CoV infection. 10.1128/mbio.02415-22.2 XBP1 is spliced in MHV-infected cells. (A) Schematic of method and primer design used to quantify %XBP1. (B) A549-MHVR cells were mock infected or infected with MHV (MOI, 0.1). Total RNA was harvested at 48 h postinfection. Relative %XBP1s, XBP1s, total XBP1, and DNAJB9 mRNA expression were quantified by RT-qPCR. CT values were normalized to 18S rRNA and expressed as the fold change over the mock control displayed as 2−Δ(ΔCT). Technical replicates were averaged, and the mean for each biological replicate (n = 2) is displayed ± SD (error bars). Download FIG S2, PDF file, 0.05 MB. Copyright © 2022 Nguyen et al. 2022 Nguyen et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. Surprisingly, despite the observed IRE1α autophosphorylation following SARS-CoV-2 infection, there was no significant upregulation of XBP1s mRNA in A549-ACE2 cells up to 52 hpi (Fig. 3C and G). Similarly, DNAJB9 expression levels were unchanged at all time points observed with SARS-CoV-2 (Fig. 3C). To confirm that this effect is not limited to A549 cells, we measured XBP1 mRNA splicing in MERS-CoV- and SARS-CoV-2-infected Calu-3 cells. Again, infection with MERS-CoV, but not SARS-CoV-2, significantly induced XBP1s and its downstream effector DNAJB9 (Fig. 3D and H). In agreement with these results, OC43, but not SARS-CoV-2, infection induced XBP1s protein levels (Fig. 3I and J).
To determine how different coronaviruses impact the UPR at the transcriptional level, we performed RNA-sequencing of A549-DPP4 cells infected with MERS-CoV for 24 and 36 h. We compared the results to published RNA sequencing (RNA-seq) data sets (35, 37) of MHV infection of murine bone marrow-derived macrophages (BMDM) or SARS-CoV-2 infection of A549-ACE2, normal human bronchial epithelial (NHBE) cells, and Calu-3 cell lines. In agreement with our IRE1α activation results, Ingenuity Pathway Analysis (IPA) predicted activation of the UPR and ER stress pathways by MERS-CoV and MHV (Fig. 4A). In contrast, SARS-CoV-2 consistently showed little to no activation of the UPR and ER stress pathway across different multiplicity of infection (MOI) conditions and cell lines. To confirm the results of the gel electrophoresis splicing assays for XBP1 mRNA that distinguished SARS-CoV-2 infection from that of the other betacoronaviruses (Fig. 3), we further utilized the RNA-seq results to quantitatively measure XBP1 mRNA splicing by these coronaviruses. Through RNA-seq, we visualized both the unspliced and spliced XBP1 mRNA reads based on whether they contain the 26-nucleotide nonconventional intron that is removed as a result of RNase activity of IRE1α as previously described (38) (Fig. 4B and C). MERS-CoV infection resulted in significant XBP1 mRNA splicing, in contrast to no difference detected in SARS-CoV-2-infected versus mock-infected cells (Fig. 4B and C). We further quantified total XBP1 spliced versus unspliced reads, which consistently showed a substantial increase in the percent expression of the XBP1s reads when normalized to total XBP1 reads for MERS-CoV at both 24 and 36 hpi but not for SARS-CoV-2-infected cells (Fig. 4D and E). This was consistent with significant upregulation of DNAJB9 and total XBP1 during infection with MERS-CoV but not SARS-CoV-2 (Fig. 4F to I).
To confirm our results in a more physiologically relevant cell, we infected iPSC-derived type II alveolar (iAT2) cells. We employed the SPC2 line, which expresses tdTomato from the surfactant protein-C (SFTPC) locus as an AT2 marker, which we have previously used to characterize innate immune responses to SARS-CoV-2 infection (3). Type II alveolar cells are a major target during both MERS-CoV and SARS-CoV-2 infection in humans, and their destruction may be a contributing factor to lung pathogenesis in severe cases (39, 40). Both MERS-CoV and SARS-CoV-2 replicate in these cells and release infectious virus as quantified by plaque assay (Fig. 5A). Notably, MERS-CoV replicated to higher titers than SARS-CoV-2 in these lung-derived cells. This complements our previous findings that SARS-CoV-2 replicates more efficiently than MERS-CoV in upper respiratory-derived primary nasal cells (3) and may suggest that MERS-CoV is better adapted to replicate within the lower respiratory tract while SARS-CoV-2 replicates more efficiently in the upper airway. Despite this difference in replication, both viruses were observed to induce p-IRE1α over the course of infection (Fig. 5B). In agreement with our results in A549 and Calu-3 cells, SARS-CoV-2 failed to induce XBP1 splicing in iAT2 cells, as measured by reverse transcription quantitative PCR (RT-qPCR) (Fig. 5C). In contrast, MERS-CoV induced XBP1 splicing, albeit to a lower extent than in immortalized cell lines. Lastly, we visualized XBP1 splicing using reverse transcriptase PCR (RT-PCR) and agarose gel electrophoresis (Fig. 5D). Again, our data indicate that SARS-CoV-2 fails to induce XBP1 splicing at either 24 or 48 hpi in iAT2 cells, despite inducing p-IRE1α. MERS-CoV, however, induced increasing XBP1 splicing over the course of infection, matching the results in A549 and Calu-3 cells (Fig. 2 and 3). Overall, these results indicate that both SARS-CoV-2 and MERS-CoV induce ER stress as evidenced by IRE1α phosphorylation during infection of primary iAT2 cells, but only MERS-CoV induces the downstream effects of active IRE1α RNase.
We then tested whether SARS-CoV-2 actively inhibits splicing of XBP1 induced by the N-linked glycosylation inhibitor tunicamycin (TM), a common agent used to chemically induce ER stress. To do so, A549-ACE2 cells were either mock infected or infected with SARS-CoV-2 or OC43 for 24 h and then treated with TM for 6 h prior to analysis. Interestingly, while SARS-CoV-2 infection did not completely prevent XBP1 splicing induced by TM, it led to significantly lower XBP1 splicing levels compared with mock infected cells (Fig. 6A). Furthermore, this inhibition is not due to a reduction in phosphorylation of IRE1 (Fig. S3A). In contrast, OC43 increased XBP1 splicing at all tested concentrations of TM (Fig. 6B). This suggests that SARS-CoV-2 actively inhibits activation of the IRE1α RNase. 10.1128/mbio.02415-22.3 Validation of IRE1α and XBP1 KO cell lines using CRISPR/Cas9. (A) A549-ACE2 cells were mock infected or infected (in triplicate) with SARS-CoV-2 (MOI, 3) for 24 h prior to treatment with low doses of tunicamycin (100 to 175 ng/mL) for 6 h. Total protein was harvested and used to quantify phospho-IRE1α and total IRE1α protein levels by Western blotting. Data shown are from one representative experiment of at least three independent experiments. (B to D) A549 cells expressing the indicated viral receptors subjected to CRISPR/Cas9 editing using different guide RNAs targeting IRE1α were immunoblotted for IRE1α protein to assess KO efficiency. (E) CRISPR/Cas-9 gene-edited IRE1α KO A549-ACE2 cells were treated with tunicamycin (500 ng/mL) or DMSO for 6 h. Total RNA was harvested, and %XBP1 was quantified by RT-qPCR. Technical replicates were averaged, and the means for each replicate displayed. Data shown are one representative experiment from at least three independent experiments. (F) CRISPR/Cas9 gene-edited IRE1α KO A549-ACE2 (guide 3) or control A549-ACE2 were treated with tunicamycin (TM; 1 μg/mL) for 8 h. Total RNA was harvested, reverse transcribed, and amplified for XBP1. XBP1 cDNA product was assayed on an agarose gel to visualize splicing. (G) Control or IRE1α KO A549-DDP4 cells were infected with MERS-CoV (MOI, 1). At the indicated time points, total RNA was collected. RT-PCR was performed using primers crossing the XBP1 splicing site. The product was analyzed on an agarose gel to visualize XBP1 splicing. (H) CRISPR/Cas9 gene-edited control or XBP1 KO A549-ACE2 was treated with DMSO or tunicamycin (TM; 1 μg/mL) for 6 h. Lysates were then immunblotted for XBP1s to confirm KO efficiency. Download FIG S3, PDF file, 0.6 MB. Copyright © 2022 Nguyen et al. 2022 Nguyen et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
Given the presumed importance of IRE1α/XBP1s to expand the ER and maintain protein folding during viral replication, and the interesting differences we observed between SARS-CoV-2 and the other betacoronaviruses, we next explored the consequences of its inhibition on the replication of each virus. To determine whether IRE1α activity is required for replication and propagation of MHV, OC43, MERS-CoV, or SARS-CoV-2, we utilized CRISPR/Cas9 gene editing to knock out IRE1α in A549 cell lines expressing receptors for each coronavirus (Fig. S3B to G). Surprisingly, we did not observe any significant differences in the capability of all tested coronaviruses to replicate in cells lacking IRE1α (Fig. 6C to F). These results suggest IRE1α is neither essential nor inhibitory for coronavirus replication in these cells. Since SARS-CoV-2 does not lead to IRE1α-mediated XBP1 splicing, we also tested replication of SARS-CoV-2 and OC43 in XBP1 KO cells (Fig. 6C and D and Fig. S3H). Consistently, there was no detectable effect of XBP1 KO on SARS-CoV-2 or OC43 replication in A549-ACE2 cells. Together, these results demonstrate that none of the coronaviruses tested require the activation IRE1α/XBP1 pathway for optimal replication.
To gain insight into the role of IRE1α in regulating betacoronaviruses, we conducted RNA-seq analysis of sg control or IRE1α knockout A549-ACE2 cells infected with either SARS-CoV-2 or OC43 compared to mock-infected cells. Infections of A549-ACE2 cells were carried out at 33°C to enable direct comparison of the two viruses (OC43 replication is significantly more robust at 33°C compared to 37°C [41], while SARS-CoV-2 replicates to a similar extent at both temperatures [Fig. S4A]). Principal-component analysis (PCA) showed a modest change in cellular gene expression upon OC43 infection of wild-type cells relative to SARS-CoV-2, which showed a robust alteration in gene expression (Fig. 7A). In contrast to uninfected or OC43-infected cells, loss of IRE1α significantly impacted host gene expression in SARS-CoV-2-infected A549 cells (Fig. 7A and B). Clustering analysis of RNA-seq data revealed 6 distinct clusters altered upon loss of IRE1α related to key cellular functions, including chromatin organization (cluster 1), mRNA metabolism and processing (cluster 2), and protein translation (cluster 3) (Fig. 7B and Fig. S5A). Detailed analysis of the IRE1α-mediated UPR pathway confirms activation by OC43 infection that is inhibited upon loss of IRE1α (Fig. 7C and Fig. S4B to E). In contrast, minimal change in this pathway was observed in SARS-CoV-2-infected cells, consistent with our previous results in this study. Loss of IRE1α also appears to alter other elements of the UPR in SARS-CoV-2-infected cells, including some genes in the PERK and ATF6 pathways (Fig. S6), which may reflect compensatory effects on the UPR in an attempt to control proteostasis in the absence of IREα (42–44). Strikingly, we observed significantly lower induction of some IFN-stimulated genes (ISGs) during SARS-CoV-2 infection of IRE1α KO cells (Fig. 7D and Fig. S4F and S5B). We have previously reported that SARS-CoV-2 induces type I and type III IFN signaling and ISGs in multiple cell types (3). Interestingly, OC43 infection did not induce notable IFN or ISG responses with or without IRE1α expression, so we were unable to make the same observations with this virus (Fig. 7D). To confirm these results, we performed RT-qPCR on representative IFN genes and ISGs genes that we have previously reported to be upregulated during SARS-CoV-2 infection (3). Consistent with our RNA-seq data, we observed significantly lower induction of ISGs such as OAS2, MX1, and IFIT1 during SARS-CoV-2 infection of cells lacking IRE1α expression at both 37°C (Fig. 7E) and 33°C (Fig. S4F). These data suggest that IRE1α may play a role in augmenting IFN signaling, while not being necessary for ISG induction, in SARS-CoV-2-infected cells. Our data taken together lead us to propose the model shown in Fig. 8. 10.1128/mbio.02415-22.4 IRE1α promotes the induction of IFN-stimulated genes upon SARS-CoV-2 infection. (A) Infection of CRISPR/Cas9-edited IRE1α KO A549-ACE2 cells with OC43 and SARS-CoV-2 (MOI, 1) under the same culture conditions at 33°C. Experiments were performed in triplicate. At the indicated times, supernatants were collected, and infectious virus was quantified by plaque assay. Values are means ± SD (error bars). Statistical significance was determined by two-way ANOVA (ns, not significant). Data shown are from one representative of at least two independent experiments. (B) Quantification of XBP1 splicing by analyzing RNA-seq data (Fig. 7). Reads representing spliced or unspliced XBP1 mRNA were identified based on the presence or absence of the 26-nucleotide intron and quantified. The percentage of XBP1 spliced reads was then plotted. Values are means ± SD (error bars). Statistical significance was determined by ordinary one-way ANOVA. (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, not significant, adjusted after Tukey’s multiple comparisons test). (C and D) Gene set enrichment analysis (GSEA) of IRE1α-mediated unfolded protein response genes with normalized enrichment score (NES) and P values compared between IRE1α KO and control cells infected with OC43 (C) or SARS-CoV-2 (D). (E) GSEA of genes that belong to GO term response to type I interferon (left) or response to interferon alpha (right) compared between IRE1α KO and control cells infected with SARS-CoV-2. (F) Infection of IRE1α KO or control A549-ACE2 with SARS-CoV-2 (MOI, 1) at 33°C. At the indicated times post infection, total RNA was collected and gene expression was quantified by RT-qPCR. CT values were normalized to 18S rRNA and expressed as the fold change over the mock control displayed as 2−Δ(ΔCT). Technical replicates were averaged, and the means for each replicate are displayed as ±SD (error bars). Statistical significance (infected compared to mock) was determined by ordinary one-way ANOVA (*, P < 0.05). Download FIG S4, PDF file, 1.2 MB. Copyright © 2022 Nguyen et al. 2022 Nguyen et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. 10.1128/mbio.02415-22.5 Metascape analysis of SARS-CoV-2 and OC43 infection RNA-seq data. (A) Metascape analyses of genes from six clusters (Fig. 7B). GO terms and KEGG pathways (hsa) are shown with –log 10 P values. (B) Ingenuity-generated interferon signaling pathway analysis comparing IRE1α KO to control cells upon SARS-CoV-2 infection from RNA-seq results (Fig. 7). Upregulated genes (red), downregulated genes (green), or no significant differential expression genes (gray) are shown with color intensity corresponding to log 2 fold-change values from RNA-seq data. Download FIG S5, PDF file, 1.3 MB. Copyright © 2022 Nguyen et al. 2022 Nguyen et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. 10.1128/mbio.02415-22.6 Transcriptomic changes in the host canonical pathway of unfolded protein response upon SARS-CoV-2 and OC43 infection. (A to C) Heatmap of normalized expression levels from RNA-seq (Fig. 7) of genes from the canonical pathway of the UPR (A), PERK branch of UPR (B), or ATF6 branch of UPR (C). Download FIG S6, PDF file, 0.2 MB. Copyright © 2022 Nguyen et al. 2022 Nguyen et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
Human respiratory betacoronaviruses initiate infection in the upper respiratory tract and have the potential to cause life-threatening pneumonia as a result of infection and inflammation of the lower respiratory tract. The host response to severe infection with coronaviruses is associated with marked dysfunction in the distal lung (alveolar) epithelium, which includes disruption of barrier function, dysregulated immune responses, transcriptomic reprogramming to a transitional cell state, and senescence (45, 46). To better understand the host epithelial response to coronavirus infection, we systematically compared the activation of the IRE1α/XBP1 pathway of the UPR during infection with betacoronaviruses in lung-derived A549 and Calu-3 cells lines and iPSC-derived AT2 cells. We employed three human viruses, each from a different betacoronavirus subgenus, OC43 (embeco), SARS-CoV-2 (sarbeco) and MERS-CoV (merbeco), and included the murine coronavirus MHV, a model embecovirus. We found a striking difference between the host response to SARS-CoV-2 and the other three viruses. OC43, MHV, and MERS-CoV all activated the canonical IRE1α/XBP1 pathway in both A549 and Calu-3 cell lines as evidenced by phosphorylation of IRE1α (Fig. 2), XBP1 mRNA splicing (Fig. 3 and 4) and induction of DNAJB9 (Fig. 3), a transcriptional target of XBP1s. Additionally, MERS-CoV was observed to induce IRE1α/XBP1 activation in iAT2 cells (Fig. 5). In contrast, while SARS-CoV-2 also promoted autophosphorylation of IRE1α, there was no evidence of XBP1s, indicating that the pathway was only partially activated and suggesting that the IRE1α kinase was active while the XBP1 splicing RNase activity was not. The differential splicing of XBP1 mRNA during SARS-CoV-2 and MERS-CoV infection was also observed in iPSC-derived AT2 cells, confirming the results in a more physiologically relevant system (Fig. 5). The difference among these viruses is surprising, as all of them encode highly conserved replicase and structural proteins that promote ER membrane rearrangements and challenge the ER folding capacity, respectively (32). We had originally hypothesized that these conserved genes would induce similar stress on the ER and lead to UPR activation. Instead, our data suggest that that SARS-CoV-2 actively prevents XBP1 splicing (Fig. 6A and B). Consistent with this idea, a recombinant SARS-CoV lacking the E protein (rSARS-CoV-ΔE) was reported to induce more XBP1 splicing as well as induction of UPR genes compared to parental wild-type virus (47). To investigate the importance of IRE1α for coronavirus replication, we evaluated replication of each of the betacoronaviruses in IRE1α KO A549 cells compared to parental wild-type cells. In contrast to influenza (48), all of the betacoronaviruses examined were able to replicate efficiently in the absence of IRE1α signaling, consistent with a previous report of the gammacoronavirus IBV (31). While we did observe a decrease in OC43 and SARS-CoV-2 nucleocapsid expression following KIRA8 treatment (Fig. 2A and F), the similar levels of replication of all the viruses in IRE1α KO cells and parental cells (Fig. 6C to F) suggest that this is due to off-target effects of KIRA8 rather than IRE1α inhibition limiting virus replication. This raises interesting possibilities for the role of IRE1α during coronavirus infection. As previously stated, IRE1α can produce both cytoprotective (through XBP1s) and destructive responses (via RIDD and JNK/p38 signaling) depending on the extent of the encountered stress. It seems likely that coronavirus infection would induce extensive and prolonged ER stress, which may push IRE1α beyond the initial pro-recovery responses and toward a pro-apoptotic response. Indeed, our data reveal that, at least with MERS-CoV and SARS-CoV-2 infection, IRE1α phosphorylation is readily detectable by 24 hpi and remains steady throughout the course of infection (Fig. S1A and B). Additionally, unlike what has been observed with chemically induced ER stress (36, 49), IRE1α phosphorylation does not appear to attenuate at any point during coronavirus infection, again suggesting a hyperactive and destructive outcome. As stated above, destruction of cells, in particular, AT2 cells in the lung, may contribute to pathogenesis during coronavirus infection. However, SARS-CoV-2 appears to limit the downstream consequences of IRE1α activation, most notably, XBP1 splicing via its RNase activity, and thus may be protected from this destructive phenotype. MERS-CoV may induce apoptosis redundantly in the UPR, as it has been reported that MERS-CoV induces and benefits from apoptosis mediated by the PERK arm of the UPR (27, 50). To further probe the impact of IRE1α signaling on host gene expression following coronavirus infection, we performed RNA-seq analysis of sg control or IRE1α knockout A549-ACE2 cells infected with either SARS-CoV-2 or OC43. IRE1α deletion significantly reduced the expression of genes downstream of XBP1s during OC43 infection, as expected, with otherwise only modest changes in overall gene expression. In contrast, genetic ablation of IRE1α significantly impacted host gene expression in SARS-CoV-2-infected A549 cells. The two most dramatic effects that appear to be specific to SARS-CoV-2 relate to chromatin organization and protein folding and transport. Effects on mRNA metabolism and processing are also observed for SARS-CoV-2 and, more modestly, for OC43. Finally, protein translation is downregulated in both OC43 and SARS-CoV-2-infected cells but, in the latter case, occurs primarily upon loss of IRE1α. Taken together, these results suggest that IRE1α plays a key role in mediating changes in host cell gene transcription and protein production caused by SARS-CoV-2. We found here that deletion of IRE1α blunted the induction of some but not all ISGs by SARS-CoV-2 infection. In contrast, OC43 was not observed to induce significant levels of IFN or ISG mRNAs in either WT or IRE1α KO cells. The mechanism by which loss of IRE1α activity during SARS-CoV-2 infection dampens the induction of interferon signaling remains to be determined. It has been reported that the UPR can precede and prime innate immune signaling in flavivirus-infected cells (51). XBP1s has been found upstream of IFNα and IFNβ transcription and may work through binding upstream cis-acting enhancer elements (52, 53). Moreover, XBP1s can directly bind and transcriptionally activate interleukin-6 (IL-6), tumor necrosis factor α (TNF-α), and other inflammatory cytokines (54). It is possible that a low level of background XBP1 splicing may occur during SARS-CoV-2 infection, which could contribute to these responses. Independent of its RNase activity, the autophosphorylated cytoplasmic domain of IRE1α can oligomerize and serve as a scaffold that recruits TRAF2, JNK, ASK, Nck, and other molecules that can lead to varied signaling outputs (55, 56). Therefore, the ability of SARS-CoV-2 to prevent full IRE1α activation might dampen inflammatory signaling and prevent detection and elimination by the immune system in an intact organism. However, it is important to note that the diminution of ISG expression in the absence of IRE1α is variable among ISGs, and SARS-CoV-2 still induces IFN and IFN signaling to a greater extent than OC43 in IRE1α KO cells. We speculate that SARS-CoV-2 has adapted to tolerate a low level of IFN signaling as well as protein kinase R (PKR) and oligoadenylate RNase L (OAS/RNase L) activation, and the reduced ISG expression in the absence of IRE1α does not have enough of an effect to promote increased replication. This is consistent with our finding that knockout of mitochondrial antiviral signaling protein (MAVS) from A549 cells, resulting in minimal IFN expression and ISG signaling, does not promote increased SARS-CoV-2 replication (3). Thus, the significance of IRE1α-dependent IFN signaling is not clear and will be a subject of future investigation. Overall, despite the lack of apparent virus replication defects with IRE1α deficiency, further characterization of the repertoire of betacoronavirus-induced IRE1α signaling is warranted, including contributions to cytokine production, apoptosis, and proinflammatory responses. While we initially investigated this pathway from the perspective of the impact on virus replication, future studies should examine effects of IRE1α activation on the host, including inflammation and cell death through the JNK and p38 mitogen-activated protein kinase (MAPK) signaling scaffolded by IRE1α (22) and/or RIDD, as a consequence of prolonged IRE1α activation (17, 57). These responses could be particularly important in AT2 cells, which must rely on the UPR to maintain proteostasis in the face of the challenge from the biosynthesis and secretion of surfactant proteins (58). Dysregulation of these responses by coronavirus infection could promote AT2 cell reprogramming, epithelial apoptosis, alteration of surfactant components in alveoli, and the rampant inflammation associated with severe coronavirus infection (59–61). Finally, the UPR response is complex and made up of the PERK and ATF6 pathways in addition to IRE1α, and signals from all three of these pathways almost certainly integrate into the final outcome of an infected cell. Indeed, changes in the PERK and ATF6 pathways may compensate for the IRE1α deficiency in the KO cells and explain the absence of an effect on replication of any of the betacoronaviruses under study. We recently reported that SARS-CoV-2 and MERS-CoV also diverge in their activation and antagonism of the dsRNA-induced host cell innate immune responses, another early innate response to viruses (3). While MERS-CoV actively antagonizes type I and type III interferon production and signaling, the oligoadenylate RNase L (OAS/RNase L) system and the PKR pathway, SARS-CoV-2 activates OAS/RNase L and PKR and induces a low level of IFN and ISG expression (3, 4) in A549 and Calu-3 respiratory tract-derived cells. Here, we observed that OC43 infection did not lead to the induction of IFN or ISGs (Fig. 7D), and we have shown previously that OC43-encoded accessory protein NS2 antagonizes activation of the OAS/RNase L pathway (62). Activation of these pathways during MERS-CoV mutant infection significantly reduces virus replication (63), while SARS-CoV-2 can tolerate the innate responses activated during infection (3). Considering the differences we have observed between betacoronaviruses with innate immune responses and now IRE1α activation and signaling, it is striking that MERS-CoV and SARS-CoV-2 are reciprocal in what they activate and antagonize. To optimize replication, coronaviruses must likely strike a balance in the cellular responses they antagonize, tolerate, or benefit from. Supporting this, our data suggest that IRE1α influences ISG induction during infection. It is intriguing to consider if MERS-CoV tolerates this by antagonizing IFN and ISG induction, while SARS-CoV-2 instead limits IRE1α activity. Future studies should examine the synergy between innate immune responses and the UPR during coronavirus infection and how perturbations on one side may change viral replicative capacity, tropism, and spread. Understanding how signals from each one of these pathways are integrated into viral replication and cell fate decisions during coronavirus infection may illuminate new therapeutic strategies for combating emerging betacoronaviruses.
Human A549 cells (ATCC CCL-185) and its derivatives were cultured in RPMI 1640 (Gibco catalog no. 11875) supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 μg/mL streptomycin (Gibco catalog no. 15140). African green monkey kidney Vero cells (E6) (ATCC CRL-1586) and VeroCCL81 cells (ATCC CCL-81) were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco catalog no. 11965) supplemented with 10% FBS, 100 U/mL of penicillin, 100 μg/mL streptomycin, 50 μg/mL gentamicin (Gibco catalog no. 15750), 1 mM sodium pyruvate (Gibco catalog no. 11360), and 10 mM HEPES (Gibco catalog no. 15630). Human HEK 293T cells (ATCC CRL-3216) were cultured in DMEM supplemented with 10% FBS. Human Calu-3 cells (ATCC HTB-55) were cultured in DMEM supplemented with 20% FBS without antibiotics. Mouse L2 cells (64) were grown in DMEM supplemented with 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, 10 nM HEPES, 2 mM l-glutamine (Gibco catalog no. 25030081), and 2.5 μg/mL amphotericin B (Gibco catalog no. 15290). A549-DPP4 (4), A549-ACE2 (3), and A549-MHVR (4) cells were generated as described previously. A549-ACE2 cells, used in Fig. 3I and J, Fig. 4, Fig. 6, and Fig. S3 were a kind gift of Benjamin TenOever, Mt. Sinai Icahn School of Medicine. CRISPR-Cas9 knockout cell lines were generated using lentiviruses. Lentivirus stocks were generated by using lentiCRISPR v2 (Addgene) with single guide RNA (sgRNA) targeting IRE1α sequences (version 1 [V1]: CGGTCACTCACCCCGAGGCC, V2: TTCAGGAAGCGTCACTGTGC, V3: CGGTCACTCACCCCGAGGCC) or XBP1 sequence (TCGAGCCTTCTTTCGATCTC). The infected A549-ACE2 cells were polyclonally selected and maintained by culture in medium supplemented with 4 μg/mL puromycin for 1 week. iPSC (SPC2 iPSC line, clone SPC2-ST-B2, Boston University)-derived alveolar epithelial type 2 cells (iAT2) were grown and infected as previously described (3). In brief, cells were differentiated and maintained as alveolospheres embedded in 3D Matrigel in CK+DCI medium, as previously described (65). For generation of 2D alveolar cells for viral infection, alveolospheres were dispersed into single cells and then plated on precoated 1/30 Matrigel plates at a cell density of 125,000 cells/cm2 using CK+DCI medium with ROCK inhibitor for the first 48 h, and then the medium was changed to CK+DCI medium at day 3 and either mock infected or infected with MERS-CoV or SARS-CoV-2 at an MOI of 5.
SARS-CoV-2 (USA-WA1/2020) was obtained from BEI Resources, NIAID, NIH or provided by Natalia Thornburg, World Reference Center for Emerging Viruses and Arboviruses (Galveston, Texas) and propagated in VeroE6-TMPRSS2 cells. The genomic RNA was sequenced and found to be identical to that of GenBank version no. MN985325.1. Recombinant MERS-CoV was described previously (1) and propagated in VeroCCL81 cells. SARS-CoV-2 and MERS-CoV infections were performed at the University of Pennsylvania or at the Howard Taylor Ricketts Laboratory (HTRL) at Argonne National Laboratory (Lemont, IL) in biosafety level 3 (BSL-3) laboratories under BSL-3 conditions, using appropriate and approved personal protective equipment and protocols. OC43 was obtained from ATCC (VR-1558) and grown and titrated on VeroE6 cells at 33°C or on A549-mRuby cells as previously described (66). MHV-A59 (5, 67) was propagated on A549-MHVR cells or on murine 17CL-1 cells.
SARS-CoV-2 and MERS-CoV infections and plaque assays were performed as previously described (1, 5). In brief, A549 cells were seeded at 3 × 105 cells per well in a 12-well plate for infections. Calu-3 cells were seeded similarly onto rat tail collagen type I-coated plates (Corning no. 356500). Cells were washed once with phosphate-buffered saline (PBS) before being infected with virus diluted in serum-free medium—RPMI for A549 cells or DMEM for Calu-3 cells. Virus was absorbed for 1 h (A549 cells) or 2 h (Calu-3 cells) at 37°C before the cells were washed 3 times with PBS and the medium was replaced with 2% FBS RPMI (A549 cells) or 4% FBS DMEM (Calu-3 cells). At the indicated time points, 200 μL of medium was collected to quantify released virus by plaque assay and stored at −80°C. Infections for MHV growth curves were performed similarly under BSL-2 conditions. For OC43 infections, similar infection conditions and media were used; however, virus was absorbed, and the infections were incubated at 33°C rather than 37°C. Plaque assays were performed using VeroE6 cells for SARS-CoV-2 and OC43, VeroCCL81 cells for MERS-CoV, and L2 cells for MHV. SARS-CoV-2 and MERS-CoV plaque assays were performed in 12-well plates at 37°C. OC43 and MHV plaque assays were performed in 6-well plates at 33°C and 37°C, respectively. In all cases, virus was absorbed onto cells for 1 h at the indicated temperatures before overlay was added. For SARS-CoV-2, MERS-CoV, and OC43 plaque assays, a liquid overlay was used (DMEM with 2% FBS, 1× sodium pyruvate, and 0.1% agarose). A solid overlay was used for MHV plaque assays (DMEM plus 2% FBS, 1× HEPES, 1× glutamine, 1× Fungizone, and 0.7% agarose). Cell monolayers were fixed with 4% paraformaldehyde and stained with 1% crystal violet after the following incubation times: SARS-CoV-2 and MERS-CoV, 3 days; OC43, 5 days; MHV, 2 days. All plaque assays were performed in biological triplicate and technical duplicate.
KIRA8 was purchased at >98% purity from Chemveda Life Sciences India Pvt. Ltd. For use in tissue culture, KIRA8 stock solution was prepared by dissolving in dimethyl sulfoxide (DMSO). Tunicamycin (catalog no. T7765) and Tg (catalog no. T9033) were purchased at >98% purity from Sigma. For use in tissue culture, tunicamycin and TG stock solutions were prepared by dissolving in DMSO.
Cells were washed once with ice-cold PBS, and lysates were harvested at the indicated times post infection with lysis buffer (1% NP-40, 2 mM EDTA, 10% glycerol, 150 mM NaCl, 50 mM Tris HCl, pH 8.0) supplemented with protease inhibitors (Roche complete mini-EDTA-free protease inhibitor) and phosphatase inhibitors (Roche PhosStop easy pack). After 5 min, lysates were incubated on ice for 20 min and centrifuged for 20 min at 4°C, and supernatants were mixed 3:1 with 4× Laemmli sample buffer (Bio-Rad 1610747). Samples were heated at 95°C for 5 min and then separated on SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes. Blots were blocked with 5% nonfat milk or 5% bovine serum albumin (BSA) and probed with antibodies (Table 1) diluted in the same blocking buffer. Primary antibodies were incubated overnight at 4°C or for 1 h at room temperature. All secondary antibody incubation steps were done for 1 h at room temperature. Blots were visualized using Thermo Scientific SuperSignal chemiluminescent substrates (catalog no. 34095 or 34080). The antibodies are listed in Table 1.
A549 cells expressing the MERS-CoV receptor DPP4 (4) were cultured in 10% FBS RPMI medium. At 70% cell confluence, cells were washed once with PBS before being mock infected or infected with MERS-CoV (EMC/2012) at and MOI of 1. Virus was absorbed for 1 h at 37°C in serum-free RPMI medium. After 1 h, virus was removed, cells were washed three times with PBS, and 2% FBS RPMI was added. The cells were incubated for another 24 h or 36 h and then washed once with PBS and lysed using RLT Plus lysis buffer before genomic DNA removal and total RNA extraction using the Qiagen RNeasy Plus minikit (Qiagen 74134). Three independent biological replicates were performed per experimental condition. RNA sample quality check, library construction, and sequencing were performed with GeneWiz following standard protocols. All samples were sequenced using an Illumina HiSeq sequencer to generate paired-end 150-bp reads. Read quality was assessed using FastQC v0.11.2 as described in reference 68. Raw sequencing reads from each sample were quality and adapter trimmed using BBDuk 38.73 as described in reference 69. The reads were mapped to the human genome (hg38 with Ensembl v98 annotation) using RNA STAR v2.7.1a (70). The resulting BAM files were counted with featureCounts v1.6.4 to count the number of reads for each gene (71). Differential expression between mock, 24 hpi, and 36 hpi experimental conditions were analyzed using the raw gene counts files by DESeq2 v1.22.1 (72). A PCA plot of RNA-seq samples and a normalized gene expression matrix were also generated with DESeq2. For SARS-CoV-2 and OC43 infections, ACE2-A549 sg control or IRE1 KO cells were cultured in 10% FBS RPMI to 70% confluence. Cells were washed once with PBS before being mock infected or infected with each virus at an MOI of 1 for 1 h in serum-free RPMI at 33°C. Cells were then washed three times with PBS before 2% FBS RPMI was added. At 48 hpi, cells were lysed with RLT Plus lysis buffer before genomic DNA removal and total RNA extraction using the Qiagen RNeasy Plus minikit (Qiagen 74134). Three independent biological replicates were performed per experimental condition. RNA sample quality check, library construction, and sequencing were performed by the University of Chicago Genomics Facility following standard protocols. All samples were sequenced in two runs using a NovaSeq 6000 sequencer to generate paired-end 100-bp reads. For each sample, the reads from two flow cells were combined before downstream processing. Quality and adapter trimming were performed on the raw sequencing reads using TrimGalore v0.6.3 (https://github.com/FelixKrueger/TrimGalore). The reads were mapped to the human genome (UCSC hg19 with GENCODE annotation), and the downstream analyses were performed using the same methods as described above.
RNA-seq data from Gene Expression Omnibus (GEO) no. GSE147507 (37), GSE168797 (38), and GSE144882 (35) and the data presented herein were used to compare the effects of different viruses on host ER stress response. Specifically, Ingenuity Pathway Analysis (IPA) (https://digitalinsights.qiagen.com/products-overview/discovery-insights-portfolio/analysis-and-visualization/qiagen-ipa/) was used to predict activities of related canonical pathways based on host gene expression changes following viral infection. Activation Z-scores for every virus and canonical pathway combination were plotted as a heatmap using Morpheus (https://software.broadinstitute.org/morpheus). IPA used the following q value cutoffs for each data set to perform the canonical pathway cross-comparison: Calu-3 SARS-CoV-2 MOI 2 24 h q < 0.05, NHBE SARS-CoV-2 MOI 2 24 h q < 0.1, A549-ACE2 SARS-CoV-2 MOI 0.2 24 h q < 0.1, A549-ACE2 SARS-CoV-2 MOI 2 24 h q < 0.05, A549-ACE2 SARS-CoV-2 MOI 3 24 h q < 0.01, A549-ACE2 SARS-CoV-2 MOI 1 48 h 33°C q < 0.05, A549-ACE2 OC43 MOI 1 48 h 33°C q < 0.001, A549-DPP4 MERS-CoV MOI 1 24 h q < 0.1, A549-DPP4 MERS-CoV MOI 1 36 h q < 0.01, BMDM MHV-A59 MOI 1 12 h q < 0.1 and over 1-fold up- or downregulated. These cutoffs were implemented due to the limitations set by the IPA software. IPA was also used to overlay gene expression data (log2 fold change) onto the interferon signaling pathway map (Fig. S5B).
Expression levels for genes involved in various pathways from RNA-seq data were drawn using Morpheus. For each gene, the normalized expression values of all samples were transformed by subtracting the mean and dividing by the standard deviation. The transformed gene expression values were used to generate the heatmap. For the clustering analysis of RNA-seq experiments for OC43- and SARS-CoV-2-infected A549-ACE2 cells with or without IRE1α, the top 5,000 most variable genes were selected. The normalized gene expression data were analyzed using Morpheus. K-means clustering with 6 clusters was applied to the gene expression data.
To identify themes across the 6 clusters, functional gene set enrichment analyses for the genes in each cluster were performed using Metascape (73). The following categories were selected for the enrichment analyses: GO molecular functions, GO biological processes, and KEGG pathway. Metascape analysis was performed with a minimum P value significance threshold of 0.05, a minimum overlap of 10 genes, and a minimum enrichment score of 5. Notable pathways enriched by Metascape from each cluster were summarized in a heatmap using Morpheus. GSEA v4.1.0 (74) was used to perform specific gene set enrichment analyses on Gene Ontology terms: IRE1-mediated unfolded protein response (75, 76), response to type I interferon (77), and response to interferon alpha (78) using the normalized expression data from the RNA-seq experiment for OC43- and SARS-CoV-2-infected A549-ACE2 cells with or without IRE1α.
All statistical analyses and plotting of data were performed using GraphPad Prism software. RT-qPCR data were analyzed by Student’s t test. Plaque assay data were analyzed by two-way analysis of variance (ANOVA) with multiple-comparison correction. Displayed significance is determined by the P value; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, not significant.
BAM files produced using RNA STAR were analyzed in Integrative Genomics Viewer v2.9.4 to count the number of XBP1 reads containing the alternative splicing (79). The total number of XBP1 reads was counted with featureCounts. The percentage of XBP1 alternative splicing for each sample was determined by dividing the number of alternatively spliced reads by the number of total XBP1 reads (spliced plus unspliced).
Cells were lysed with RLT Plus buffer, and total RNA was extracted using the RNeasy Plus minikit (Qiagen). RNA was reverse transcribed into cDNA with a high-capacity cDNA reverse transcriptase kit (Applied Biosystems 4387406). cDNA samples were diluted in molecular biology-grade water and amplified using specific RT-qPCR primers (see Table 2). RT-qPCR experiments were performed on a Roche LightCycler 96 instrument. SYBR green supermix was from Bio-Rad. Host gene expression displayed as the fold change over mock-infected samples was generated by first normalizing cycle threshold (CT) values to 18S rRNA to generate ΔCT values (ΔCT = CT gene of interest − CT 18S rRNA). Next, Δ (ΔCT) values were determined by subtracting the mock-infected ΔCT values from the virus-infected samples. Technical triplicates were averaged and means displayed using the equation 2–Δ(ΔCT). Primer sequences are listed in Table 2.
RT-qPCR was used to quantify the relative expression of the spliced version of XBP1 (XBP1s) by using specific pairs of primers for human alternatively spliced XBP1 and total XBP1 (primer sequences are described above) as previously described (80). The relative percentage of alternative splicing of XBP1 (%XBP1s) was indicated by calculating the ratio of signals between XBP1s and total XBP1.
Raw and processed RNA-seq data for MERS-CoV, OC43, and SARS-CoV-2 were deposited in the Gene Expression Omnibus database (GSE193169). | true | true | true |
PMC9600278 | Mingxi Jia,Shanshan Feng,Fengxi Cao,Jing Deng,Wen Li,Wangyan Zhou,Xiang Liu,Weidong Bai | Identification of EGFR-Related LINC00460/mir-338-3p/MCM4 Regulatory Axis as Diagnostic and Prognostic Biomarker of Lung Adenocarcinoma Based on Comprehensive Bioinformatics Analysis and Experimental Validation | 17-10-2022 | lung adenocarcinoma,ceRNA network,LINC00460/MCM4 axis,drug targets,prognosis | Simple Summary While the epidermal growth factor receptor (EGFR) is an important target for lung adenocarcinoma (LUAD) therapy, acquired resistance is still inevitable. A comprehensive bioinformatics analysis strongly suggested that the closely related LINC00460-mir-338-3p-MCM4 ceRNA network of EGFR plays an important role in the diagnosis and prognosis of LUAD. High expression of LINC00460 and MCM4 predicts shorter patient survival. Univariate and multivariate regression analyses demonstrated that higher expression of LINC00460 and MCM4 was significantly associated with tumor size, lymph node metastasis, distant metastasis and TNM stage. A multi-gene regulation model analysis revealed that the LINC00460 (downregulation)—mir-338-3p (upregulation))—MCM4 (downregulation) pattern significantly improved the overall survival of LUAD patients. We also verified the expression of these genes in LUAD cell lines and tumor tissues by RT-PCR and immunohistochemistry. Finally, the possible targeted drugs of MCM4 were queried through the drug database platform, hoping to solve its drug resistance problem by targeting EGFR-related genes. Abstract Background: Lung adenocarcinoma (LUAD) is one of the most aggressive and lethal tumor types and requires effective diagnostic and therapeutic targets. Though the epidermal growth factor receptor (EGFR) is an important target for LUAD therapy, acquired resistance is still inevitable. In recent years, the regulation of the EGFR by competing endogenous RNAs (ceRNAs) has been extensively studied and significant progress has been made. Therefore, we aim to find new targets for the diagnosis and treatment of LUAD by analyzing the EGFR-related ceRNA network in LUAD and expect to address the problem of EGFR resistance. Methods: We identified differentially expressed lncRNAs, miRNAs and mRNAs closely associated with the EGFR by analyzing transcriptome data from LUAD samples. Comprehensive bioinformatics analysis strongly suggests that the LINC00460—mir-338-3p—MCM4 ceRNA network plays an important role in the diagnosis and prognosis of LUAD. The effects of different patterns of the LINC00460/MCM4 axis on the overall survival of patients with LUAD were analyzed by a polygene regulation model. We also verified the expression of these genes in LUAD cell lines and tumor tissues by RT-PCR and immunohistochemistry. The functional enrichment analysis and targeted drug prediction of the MCM4 gene were performed. Results: Survival analysis indicated that high expressions of LINC00460 and MCM4 predict a shorter survival period for patients. Univariate and multivariate regression analyses demonstrated that higher expressions of LINC00460 and MCM4 were significantly associated with tumor size, lymph node metastasis, distant metastasis and TNM stage. A multi-gene regulation model analysis revealed that the LINC00460 (downregulation)—mir-338-3p (upregulation)—MCM4 (downregulation) pattern significantly improved the overall survival of LUAD patients (p = 0.0093). RT-PCR and immunohistochemical experiments confirmed our analytical results. In addition, the functional enrichment analysis indicated that MCM4-related genes were mainly enriched in the cell cycle and cell division. A functional association network analysis showed that MCM4 was closely related to the EGFR. Finally, the possible targeted drugs of MCM4 were queried through the drug database platform, hoping to solve its drug resistance problem by targeting EGFR-related genes. Conclusions: In summary, the LINC00460/MCM4 axis can be used as a potential new perspective for targeting EGFR genes in precision medicine and is expected to serve as a diagnostic, prognostic and drug target for LUAD. | Identification of EGFR-Related LINC00460/mir-338-3p/MCM4 Regulatory Axis as Diagnostic and Prognostic Biomarker of Lung Adenocarcinoma Based on Comprehensive Bioinformatics Analysis and Experimental Validation
While the epidermal growth factor receptor (EGFR) is an important target for lung adenocarcinoma (LUAD) therapy, acquired resistance is still inevitable. A comprehensive bioinformatics analysis strongly suggested that the closely related LINC00460-mir-338-3p-MCM4 ceRNA network of EGFR plays an important role in the diagnosis and prognosis of LUAD. High expression of LINC00460 and MCM4 predicts shorter patient survival. Univariate and multivariate regression analyses demonstrated that higher expression of LINC00460 and MCM4 was significantly associated with tumor size, lymph node metastasis, distant metastasis and TNM stage. A multi-gene regulation model analysis revealed that the LINC00460 (downregulation)—mir-338-3p (upregulation))—MCM4 (downregulation) pattern significantly improved the overall survival of LUAD patients. We also verified the expression of these genes in LUAD cell lines and tumor tissues by RT-PCR and immunohistochemistry. Finally, the possible targeted drugs of MCM4 were queried through the drug database platform, hoping to solve its drug resistance problem by targeting EGFR-related genes.
Background: Lung adenocarcinoma (LUAD) is one of the most aggressive and lethal tumor types and requires effective diagnostic and therapeutic targets. Though the epidermal growth factor receptor (EGFR) is an important target for LUAD therapy, acquired resistance is still inevitable. In recent years, the regulation of the EGFR by competing endogenous RNAs (ceRNAs) has been extensively studied and significant progress has been made. Therefore, we aim to find new targets for the diagnosis and treatment of LUAD by analyzing the EGFR-related ceRNA network in LUAD and expect to address the problem of EGFR resistance. Methods: We identified differentially expressed lncRNAs, miRNAs and mRNAs closely associated with the EGFR by analyzing transcriptome data from LUAD samples. Comprehensive bioinformatics analysis strongly suggests that the LINC00460—mir-338-3p—MCM4 ceRNA network plays an important role in the diagnosis and prognosis of LUAD. The effects of different patterns of the LINC00460/MCM4 axis on the overall survival of patients with LUAD were analyzed by a polygene regulation model. We also verified the expression of these genes in LUAD cell lines and tumor tissues by RT-PCR and immunohistochemistry. The functional enrichment analysis and targeted drug prediction of the MCM4 gene were performed. Results: Survival analysis indicated that high expressions of LINC00460 and MCM4 predict a shorter survival period for patients. Univariate and multivariate regression analyses demonstrated that higher expressions of LINC00460 and MCM4 were significantly associated with tumor size, lymph node metastasis, distant metastasis and TNM stage. A multi-gene regulation model analysis revealed that the LINC00460 (downregulation)—mir-338-3p (upregulation)—MCM4 (downregulation) pattern significantly improved the overall survival of LUAD patients (p = 0.0093). RT-PCR and immunohistochemical experiments confirmed our analytical results. In addition, the functional enrichment analysis indicated that MCM4-related genes were mainly enriched in the cell cycle and cell division. A functional association network analysis showed that MCM4 was closely related to the EGFR. Finally, the possible targeted drugs of MCM4 were queried through the drug database platform, hoping to solve its drug resistance problem by targeting EGFR-related genes. Conclusions: In summary, the LINC00460/MCM4 axis can be used as a potential new perspective for targeting EGFR genes in precision medicine and is expected to serve as a diagnostic, prognostic and drug target for LUAD.
Lung cancer is the most common and deadliest form of cancer. Non-small cell lung cancer (NSCLC) accounts for approximately 80% of all lung cancers and has a low five-year survival rate [1,2]. LUAD is the most common histological subtype of NSCLC, and the five-year survival rate of patients represents only 15% of them [3]. Therefore, it is still very important to determine new diagnostic, prognostic and drug resistance biomarkers and treatment targets for LUAD research. In tumor cells, epidermal growth factor receptor (EGFR) activity may be dysregulated due to mutations, increased gene copy number or protein overexpression [4,5]. Among 441 LUAD patients, 218 (49.4%) patients had wild-type EGFR, and 223 (50.6%) had mutant EGFR [6]. Similar studies have shown that approximately 60% of NSCLC cases associated with poor prognosis have EGFR overexpression or constitutive activation [7]. Overexpression or mutation of the EGFR gene greatly promotes cell growth and division in NSCLC. Several pieces of research have indicated that the overexpression of EGFR is related to the low survival rate, frequent lymph node metastasis and poor chemotherapy sensitivity of NSCLC patients [7,8]. Furthermore, most EGFR mutations in NSCLC occur in the exons of the receptor tyrosine kinase domain [9]. Although EGFR tyrosine kinase inhibitors (EGFR-TKIs), including erlotinib and gefitinib, have shown initial efficacy in 30% of NSCLC patients with EGFR mutations in the past few decades, secondary resistance often occurs in EGFR-TKIs treatment [10,11]. Currently, multiple mechanisms of secondary resistance to EGFR-TKIs, including primary or secondary T790M point mutations, human epidermal growth factor receptor 2 (HER2) amplification, mesenchymal epithelial cell transforming factor (MET) amplification or activation of bypass signaling pathways by phosphatidylinositol 3 kinase (PI3K) mutations and epithelial –mesenchymal transitions (EMT) have been clarified. However, the combination of EGFR-TKIs with platinum or other cytotoxic chemotherapeutic agents did not achieve the expected prolongation of survival in NSCLC patients, resulting in increased toxicity and side effects [12,13]. Therefore, it is of great significance to deeply study the molecular mechanism of EGFR for the diagnosis of NSCLC, the development of new drugs and the formulation of new treatment strategies to prolong the survival of patients. In recent years, the regulation of EGFR by competing endogenous RNAs (ceRNAs) has been extensively studied [14,15]. Long non-coding RNA (lncRNA) can regulate gene expression at multiple levels through epigenetic regulation, transcription regulation and post-transcriptional regulation, and then participate in a variety of biological processes [16,17,18]. In the lncRNA-miRNA-mRNA ceRNA network, lncRNA can regulate mRNA expression through sponge adsorption of miRNA [19,20,21]. Studies have shown that EGFR, mRNA and protein levels are regulated by a large number of protein-coding and noncoding RNAs, most of which are mRNAs unrelated to EGFR function, but capable of “protecting” it from “attack” by their shared miRNAs [22,23]. For example, the lncRNA SNHG16 is highly expressed in gliomas and acts as a ceRNA to regulate EGFR by sponging miR-373-3p to activate the PI3K/AKT pathway, thereby exerting oncogenic effects [24]. These EGFR-centered ceRNA regulatory networks are very important. Several studies have shown that high expression of EGFR is an important indicator of poor prognosis in LUAD [25]. High levels of EGFR expression are associated with resistance to chemotherapeutic drugs. A high expression of EGFR was associated with increased gene copy numbers, and non-small cell lung cancer patients given gefitinib were associated with significantly improved responses, lower progression rates and improved survival compared with tumors with low EGFR expression [26]. Based on the above description, we aim to find new targets for the diagnosis and treatment of LUAD by analyzing the EGFR-related ceRNA network. We conducted a systematic and comprehensive study on the RNAseq and miRNAseq data of LUAD from TCGA. A ceRNA network closely related to the EGFR was identified by expression analysis and survival analysis. The correlation between different expression patterns and overall survival (OS) was analyzed using a multi-gene regulation model. We also verified the expression of these genes in LUAD cell lines and tumor tissues by RT-PCR and immunohistochemistry. Finally, the possible targeted drugs of the target gene were queried through the drug database platform, hoping to solve its drug resistance problem by targeting EGFR-related genes.
The RNA data and clinical information of lung adenocarcinoma samples were obtained from the TCGA database (https://portal.gdc.cancer.gov/, accessed on 18 January 2021). Finally, 515 tumor samples and 57 paracancerous samples were identified as research objects by matching the patient numbers corresponding to the lncRNA, mRNA and miRNA samples.
According to the median expressed in the EGFR (median value = 6616), 515 LUAD patients were divided into EGFRlow group (EGFR < 6616, n = 257) and EGFRhigh group (EGFR ≥ 6616, n = 258). Performing the differential expression analysis in EGFRlow and EGFRhigh LUAD samples. The differentially expressed lncRNAs had thresholds of |log2FC| > 0.70 and p < 0.05, the differentially expressed miRNAs and mRNAs had thresholds of |log2FC| > 0.50 and p < 0.05.
The ceRNA network was constructed according to the method described previously [27]. Briefly: (1) Prediction of potential miRNAs for DElncRNA targeting through the mircode database (http://www.mircode.org/, accessed on 20 March 2022); (2) miRDB databases (http://www.mirdb.org/miRDB/, accessed on 20 March 2022) and the Targetscan database (http://www.targetscan.org/, accessed on 20 March 2022) were used to identify the downstream targets (mRNAs) of miRNAs; (3) The lncRNA-miRNA-mRNA ceRNA network was visualized by Cytoscape v3.7.0.
A Kaplan–Meier survival analysis displayed the correlation between the expression of DERNAs in the ceRNA network and the prognostic survival of LUAD patients. The survival status and time of LUAD patients were obtained from TCGA clinical data. Associations between candidate genes and overall survival (OS) were analyzed to determine biological prognostic markers by univariate and multivariate Cox regression. In addition, a novel analysis model analyzed the correlation between different expression combination patterns of three genes (LINC00460, mir-338-3p and MCM4). LINC00460 was divided into high expression group “L+” and low expression “L−” based on their median expression values. Similarly, mir-338-3p was classified into “m+” and “m−” groups, and MCM4 was divided into “M+” and “M−” groups. Eight combined expression patterns were ultimately obtained, including: L+/m+/M+, L+/m+/M−, L+/m−/M−, L−/m+/M−, L−/m+/M+, L+/m−/M+, L−/m−/M− and L−/m−/M+. The regulatory mechanism of the LINC00460—mir-338-3p—MCM4 axis on the prognosis and survival of LUAD patients was analyzed through a multi-gene interaction model.
Lung adenocarcinoma cell lines (A549, PC-9 and H1299) and the normal bronchial epithelial cell line (BEAS-2B) were cultured in 1640 medium (Thermo Fisher Scientific, Waltham, MA, USA) containing 10% fetal bovine serum (FBS) (Thermo Fisher Scientific, Waltham, MA, USA), 100 U/mL streptomycin and 100 U/mL penicillin (Sangon Biotech, Shanghai, China) at 37 °C and 5% CO2. RNA was extracted according to the previous experimental method and detected by RT-PCR [27]. The primer sequences were as follows: GAPDH, F: 5′-CAGGAGGCATTGCTGATGAT-3′, R: 5′-GAAGGCTGGGGCTCATTT-3′. LINC00460, F: TCGGCTAAGAGTCACCCTGGATG-3′, R: 5′- CACAGACGCCTCCCACACAATG-3′. MCM4, F: 5′- ATCTCCCTCTCAGAGACGTAG-3′ R: 5′-TGTCAGTGGTGAACTAACATCA-3′. U6, F: 5′-AGAGAAGATTAGCATGGCCCCTG-3′, R: 5′-AGTGCAGGGTCCGAGGTATT-3′. miR-338-3p, F: 5′-CGCGTCCAGCATCAGTGATT-3′, R: 5′-AGTGCAGGGTCCGAGGTATT-3′.
LUAD tumor tissue and paracancerous tissue were obtained from the Department of Hematology & Oncology, the First Hospital of Changsha. This study was approved by the ethics committee. The collection of tissue specimens was explained in detail to the patients or their families in advance, and an “informed consent form” was signed after obtaining the consent of the patients and their families. Tissues were fixed and embedded in paraffin, then cut into 5 μm thick sections and mounted on New Silane slides with a MCM4 mouse polyclonal antibody (Cat.# D260599, Sangon Biotech, Shanghai, China). Standard hematoxylin-eosin (HE) staining and immunohistochemical (IHC) were used to assess protein expression levels in tumor samples [28].
The top 200 genes associated with the MCM4 gene in LUAD were gained from GEPIA (http://gepia.cancer-pku.cn/, accessed on 28 March 2022). The Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) were analyzed by DAVID (https://david.ncifcrf.gov/tools.jsp, accessed on 28 March 2022) and visualized by “ggplot2”. The functional association network of the target gene was predicted by GeneMINIA (http://genemania.org/, accessed on 28 March 2022). The “Multiple proteins” module of the STRING database (https://string-db.org/, accessed on 29 March 2022) was used to map the top 10 protein interaction networks associated with the target genes based on the same function that the proteins co-promote.
Through cBioPortal (https://www.cbioportal.org/, accessed on 5 April 2022), the mutation rate, types and common mutation sites of MCM4 in tumors were acquired, and the relationship between MCM4 mutation and clinical prognosis of patients was explored.
We hope to address the problem of EGFR target resistance through the treatment of downstream genes of the EGFR. We performed a drug target analysis of the MCM4 gene through the HERB (http://herb.ac.cn/, accessed on 5 April 2022) and GSCAlite (http://bioinfo.life.hust.edu.cn/web/GSCALite/, accessed on 5 April 2022) databases. The expectation is that these targeted drugs may play a role in the treatment of LUAD.
Correlations between gene expression levels and tumor mutational burden (TMB) were evaluated based on the mutation data of LUAD downloaded from TCGA. We also used “Estimate” in the R package to score the stromal and immune cells of the samples, where the stromal score represents the number of stromal cells in the tumor immune microenvironment (TIME) and the immune score represents the number of immune cells in TIME; the sum of these two is the total score.
All statistical analyses were performed by GraphPad Prism 8.4.3 and SPSS version 23.0. Statistically significant differences between the two groups of data were estimated by the Mann–Whitney test and independent t-test. One-way ANOVA was used to assess statistical differences between multiple groups of data. A p-value < 0.05 was considered statistically significant.
EGFR overexpression in lung cancer tissue was found based on the Human Protein Atlas database (HPA, http://www.proteinatlas.org/, accessed on 15 January 2021) (Figure 1A and Supplementary Figure S1). In addition, the cBioPortal database (http://www.cbioportal.org/, accessed on 15 January 2021) showed that changes in EGFR gene expression in the TCGA LUAD dataset were mainly due to its amplification and missense mutations (Figure 1B). The increase in gene copy numbers was likely to be one of the main mechanisms that made a contribution to the over-regulation of the EGFR in LUAD patients (Supplementary Figure S1). Similar EGFR expression imbalances were confirmed by IHC from the HPA database (Figure 1C), and patients’ information was shown in Table S1. The EGFR was highly and differentially expressed in tumor samples between tumor-paracancer (p = 0.0301) and paired tumor-normal (p = 0.0451) samples (Figure 1D,E). In addition, some traditional prognostic factors were also analyzed (Figure 1F–J), and those results displayed a significant correlation between tumor metastasis and high expression of the EGFR (p = 0.0057) (Figure 1I). The high EGFR expression could be closely related to tumor metastasis in LUAD.
Based on the above analysis, the ceRNA network associated with the EGFR can be used as a potential prognostic model for patients of LUAD. Moreover, we must make it clear that the meaning of the expression levels in LUAD samples with EGFRhigh and EGFRlow expression groups are consistent with those in cancer and paracancerous groups. In total, 1098 DElncRNAs, 36 DEmiRNAs and 4738 DEmRNAs were screened in LUAD samples with EGFRhigh and EGFRlow expression groups. The distribution of DElncRNAs, DEmiRNAs and DEmRNAs were visualized by the volcano plot as shown in Figure 2A–C.
A total of 62 lncRNAs and their potential target 4 miRNAs were identified based on the TarBase database. Meanwhile, the target mRNAs of these 4 miRNAs were identified by TargetScan and miRDB databases. Ultimately, 376 of the 4738 DEmRNAs were identified. Finally, Cytoscape software was used to construct and visualize the EGFR-related lncRNA-miRNA-mRNA triple regulatory network in LUAD, including 62 lncRNAs, 4 miRNAs and 376 mRNAs (Figure 2D). In order to further explore the potential functions of the ceRNA network related to the EGFR, a functional enrichment analysis (including GO and KEGG) was performed on these mRNAs by Metascape (Figure 2E). In order to determine whether these RNAs were related to prognosis in LUAD, we first performed an OS analysis of LUAD patients using a Kaplan–Meier analysis and log-rank test. In total, 7 DElncRNAs, 2 DEmiRNAs and 37 DEmRNAs were found to be associated with prognosis (Figure 3). Comparing the lncRNA-miRNA and miRNA-mRNA target gene matching results from the above analysis, a survival-related ceRNA regulatory network including seven lncRNAs, two miRNAs and eight mRNAs were finally constituted (Figure 2F).
To identify ceRNA networks of significant prognostic value in LUAD, we further explored RNA expression levels in high and low EGFR expression groups as well as in tumor and adjacent normal lung tissue (Figure 4A). The results showed six upregulated (AC084083.1, ARHGEF26-AS1, VIPR1-AS1, LINC00342, MEG3, LINC00460) and one downregulated (AP002478.1) lncRNAs, two downregulated (mir-215, mir-338-3p) miRNAs, seven upregulated (MCM4, CDIP1, RAB27B, TMEM255A, LSAMP, ATP8B4, SRGAP3) and one undifferentiated (CLCN3) mRNAs in LUAD samples with EGFRhigh and EGFRlow groups (Figure 4A). In addition, the expression levels of these RNAs were confirmed in 54 (or 46) paired LUAD samples (Figure 4B), as well as in 59 (or 46) normal samples and 535 (or 521) LUAD samples (Figure S2). Then, the LINC00460 (upregulated)—mir-338-3p (downregulated)-MCM4 (upregulated) axis was finalized according to the results of the expression validation of DERNAs and survival analysis. It was found that the expression of LINC00460 was negatively correlated with the expression of mir-338-3p and positively correlated with the expression of MCM4 by expression correlation analysis (Figure 5A). Pairing between mir-338-3p and the target sites in LINC00460 and MCM4 was predicted by MiRcode and TargetScan, respectively (Figure 5B). These data indicate that LINC00460 can enhance the expression of MCM4 by sponge adsorption mir-338-3p.
We have creatively developed a new analytical model to explore the correlation between different combinations of expression patterns of three genes (LINC00460, mir-338-3p and MCM4) with LUAD patients and OS. The results indicated that the OS of patients in the pattern of L−/m+/M- was markedly improved compared to the expression pattern of L+/m−/M+ (p = 0.0093) (Figure 5C,D). Furthermore, other expression modes showed that the OS of LUAD patients in the pattern of L−/m+/M− was markedly improved compared with the pattern of L+/m+/M+ and L+/m−/M−. The results manifested that the downregulation of LINC00460 and MCM4 expression levels, or the promotion of mir-338-3p expression, could inhibit the occurrence and development of LUAD cells. These were consistent with the above analysis results, illustrating the feasibility and accuracy of our new analysis model.
To understand whether the expression levels of LINC00460 and MCM4 were influenced by clinical characteristics, the correlation was explored between LINC00460 and MCM4 expression with clinical factors. These results indicated that the expression of both LINC00460 and MCM4 were positively correlated with lymph node metastasis (LINC00460, p = 0.0006; MCM4, p = 0.0463) and TNM stage (LINC00460, p = 0.0468; MCM4, p = 0.0356) (Table 1), and MCM4 expression was also positively correlated with distant metastasis (p = 0.021). In addition, MCM4 expression values were significantly higher in men than in women (p = 0.0006) and higher in patients aged <60 than in patients aged ≥60 (p = 0.012, Table 1). We also detected the correlation of LINC00460 and MCM4 with the overall survival and TNM stage of LUAD patients in the GEPIA database. The results revealed that the high expression of LINC00460 and MCM4 was significantly related to the poor prognosis of LUAD patients (p < 0.01, Figure S3), which is consistent with our analysis results. Furthermore, univariate and multivariate Cox regression analyses were used to find out the correlation between clinical features and OS. In the univariate Cox regression analysis model of LINC00460 and MCM4, clinical prognostic factors such as TNM stage, tumor size and lymph node metastasis of LUAD patients were closely related to OS (p < 0.05; Table 2 and Table 3). Importantly, both LINC00460 (HR = 1.383, p = 0.043) and MCM4 (HR = 1.608, p = 0.003) over-expression significantly related to a worse prognosis (Table 2 and Table 3). The multivariate Cox regression analysis of MCM4 indicated that tumor size (HR = 1.550, p = 0.012), lymph node metastasis (HR = 1.974, p < 0.0001) and high expression of MCM4 (HR = 1.459, p = 0.019) were closely related to OS in LUAD patients (Table 3). In summary, MCM4 may serve as an important prognostic factor for patients with LUAD.
The expression levels of the LINC00460-miR-338-MCM4 regulatory axis in lung adenocarcinoma cells (A549, PC-9 and H1299) and normal human bronchial epithelial cells (BEAS-2B) were detected by RT-PCR (Figure 6). The expression levels of LINC00460 and MCM4 were significantly up-regulated in A549, PC-9 and H1299 relative to BEAS-2B cells, while the expression of miR-338 was significantly down-regulated (p < 0.05). This confirmed the accuracy of the bioinformatics analysis results. In addition, the expression of the MCM4 protein in lung adenocarcinoma tumors and paracancerous tissue was also examined by IHC (Figure 6D), and the results indicated that MCM4 was significantly highly expressed in LUAD tumor tissues.
In order to further explore the possible function of MCM4 in LUAD, GO and KEGG enrichment analyses were performed on the top 200 MCM4-related genes in LUAD (Figure 7). The enrichment terms associated with MCM4 were “p53 signaling pathway” and “cell cycle”. In addition, a GO enrichment analysis indicated that MCM4 mainly enriched in “cell division”, “mitotic nuclear division”, “nucleoplasm” and “protein binding”. The functional association network of MCM4 and EGFR was predicted by the GeneMINIA tool and 20 potential target genes were shown (Figure 7E,F). MCM2, MCM3, MCM5, MCM6, MCM7 and MCM4 belong to the MCM family and are closely related to the EGFR. MCM proteins are essential proteins for initiation and elongation steps during DNA replication in organisms. This suggests that MCM4 may affect the occurrence and development of LUAD by regulating processes such as DNA replication and the cell cycle.
An HERB database analysis identified 4 drugs with MCM4 as a potential target (Figure 8A). Furthermore, we used the GSCALite database and assessed the correlation of drug sensitivity with MCM4 according to the Cancer Therapy Response Portal (CTRP) and genomics of drug sensitivity in cancer (GDSC). The CTRP illustrated that MCM4 was resistant to Docetaxel and sensitive to 17 drugs including 5-Fluorouracil (5-fluorouracil), PHA-793887 and KIN001-102 (Figure 8B). Figure 8C shows that the high expression level of MCM4 was also sensitive to the other 138 drugs. These drugs may target MCM4 and play a role in the treatment of LUAD. In addition, the cBioPortal tool was used to analyze the mutational signature of MCM4. The OncoPrint plot showed the amplification of MCM4 genomic in the TCGA LUAD dataset in Figure S4A. A significant association was observed between MCM4 expression and copy number in LUAD samples (Figure S4D). Therefore, these data indicated that the aberrant expression of MCM4 was mainly due to copy number changes in LUAD rather than gene mutations. Figure S4B,C illustrated several highly mutated sites in the MCM4 gene and their positions in the three-dimensional structure. In addition, it can also be seen that the most mutated forms of MCM4 in pan-cancer cells are also amplified and missense mutations (Figure S4E). This suggested that the pro-oncogenic mechanism of MCM4 in LUAD was similar to that of other pan-cancer mechanisms.
An iammune correlation analysis displayed that the expression levels of six immune markers of neutrophils (CD66b) and dendritic cells (HLA-DPB1, HLA-DQB1, HLA-DRA, HLA-DPA1 and BDCA-1) have significant negative correlations with MCM4 expression in LUAD. In addition, MCM4 expressions have significant negative correlations with the expression levels of four immune markers of three immune cells, including the natural killer cell (KIR2DL4), Th1 (STAT1, IFN-γ) and T cell exhaustion (GZMB) (Table S2). We also verified the relationship between MCM4 expression levels in LUAD and the above-mentioned markers using the GEPIA database and obtained similar results (Table S3). Tumor mutation burden (TMB) represents the number of mutations in tumor cell genes per million genes in a patient sample. The correlation results between TMB and MCM4 showed that the expression level of MCM4 correlated with TMB (Figure S5A). TIME values for LUAD samples were assessed and compared with gene expression and patient OS (Figure S5B–D). The results showed that there was a significant negative correlation between MCM4 expression and stromal score, immune score and total score. The lower the content of stromal cells and immune cells in the tumor environment, the higher the purity of the tumor. These results demonstrated that the LINC00460/MCM4 axis may regulate the level of tumor infiltrating immune cells in LUAD, which in turn has an impact on clinical outcomes.
The EGFR gene regulates many physiological processes and induces important mechanisms associated with cancer [29,30]. Previous studies have found that a high expression and high copy number of the EGFR were demonstrated in 10–30% of NSCLC patients [31]. Earlier studies suggested that constitutive activation of the EGFR signaling pathway in tumor tissues may be initiated by EGFR gene amplification or triggered by EGFR mutations [5,32,33]. In contrast, Li et al. found that EGFR over-expression was mainly closely related to amplification but statistically independent of EGFR mutations in lung adenocarcinoma [34]. The specific mechanism of EGFR action in tumors has not been clarified and further studies are needed. In this study, we obtained the ceRNA regulatory network associated with EGFR expression by computer analysis. Comprehensive bioinformatics analysis strongly suggests that the LINC00460-mir-338-3p-MCM4 ceRNA network plays an important role in the diagnosis and prognosis of LUAD. The analysis results of LINC00460, mir-338-3p and MCM4 were basically consistent with the existing research reports [35,36]. LncRNAs may serve as potential predictive and prognostic markers for EGFR resistance in LUAD, as they are involved in modulating chemosensitivity, radiosensitivity and sensitivity to EGFR-targeted therapy through multiple mechanisms [37]. Studies have revealed that LINC00460 is significantly upregulated in NSCLC and promotes the metastasis and invasion of lung cancer cells by inducing epithelial–mesenchymal transformation [35]. The expression of LINC00460 is up-regulated in gefitinib-resistant NSCLC tissues and cells and is closely associated with advanced tumor stage and poor clinical prognosis. More interestingly, LINC00460 promoted EGFR expression by sponging miR-769-5p, thereby promoting the resistance of NSCLC cells to gefitinib [38]. LINC00460 may be a novel prognostic and therapeutic target for LUAD. miR-338-3p was poorly expressed in NSCLC tissue relative to adjacent noncancerous tissue. LUAD cell proliferation, migration and invasion were inhibited by miR-338-3p overexpression. miR-338-3p directly targets Neuropilin 1 (NRP1) and plays a role in enhancing drug sensitivity in EGFR wild-type NSCLC patients [39]. The mini chromosome maintenance (MCM) protein 2–7 forms the complex necessary for DNA replication to begin [40]. Kikuchi et al. [36] found that the expression level of MCM4 in NSCLC cancer cells was higher than in adjacent normal lung cells (p < 0.001), and MCM4 might play a significant role in the proliferation of NSCLC cells. However, the mechanism of action between MCM4 and EGFR has not been clearly reported. In this research, the univariate Cox regression analysis of LINC00460 and MCM4 indicated that TNM stage, tumor size and lymph node metastasis were closely related to the OS of LUAD patients. Importantly, both LINC00460 and MCM4 over-expression significantly related to a worse prognosis. A multivariate Cox regression analysis of MCM4 showed that tumor size, lymph node metastasis and MCM4 high expression were also associated with OS in LUAD patients. A multi-gene regulation model analysis revealed that the L−/m+/M− pattern significantly improved the OS compared to the L+/m−/M+ expression pattern in LUAD patients (p = 0.0093). This suggests that inhibiting the expression of LINC00460 and MCM4 or upregulating the expression of mir-338-3p can both prevent tumor progression and improve the prognosis of LUAD patients. In addition, the approach of this study provides a new idea to address tumor drug resistance. The research has indicated that RAS dysregulation and the resulting signaling dysregulation account for one-third of all human cancers, and mutations in RAS are usually related to treatment resistance and poor prognosis [41,42]. As a key cancer driver, RAS has always been the focus of an intensive search for therapeutic approaches. So far, however, no effective RAS inhibitors have been approved for clinical use. Recently, the clinical results of KRAS G12C inhibitors have sparked excitement in the scientific community [43,44]. Nevertheless, acquired drug resistance may limit the efficacy of inhibitors, indicating that combination therapy may be required [45]. By analyzing RAS-related dysregulated genes and constructing a RAS-centered ceRNA network, oncogenic RAS and its downstream signaling and metabolic programs can be more effectively targeted. An accurate understanding of the coordinated interactions between RAS and other genes in the associated ceRNA network will be very important for developing novel targeted therapies for RAS-driven cancers. In conclusion, we have determined that the ceRNA-based LINC00460/MCM4 axis may be a potential therapeutic target and prognostic biomarker of LUAD, but there are still some limitations. First, most of our study data were obtained from the TCGA database, and some of the analysis results may be biased and need to be validated by further clinical trials. Secondly, the related mechanisms of the LINC00460/MCM4 axis and EGFR gene in LUAD need to be further investigated experimentally. Finally, the specific mechanism of action between the LINC00460/MCM4 pathway and EGFR requires further experimental studies. Despite these deficiencies, a LINC00460—mir-338-3p—MCM4 regulatory network was identified by comprehensive analysis of the EGFR and ceRNA, which is expected to be an effective diagnostic and therapeutic target.
A detailed comprehensive analysis of the EGFR and ceRNA was performed, indicating that LINC00460/MCM4 can effectively predict the survival outcome of patients with LUAD and can hopefully be an effective procedure for diagnosis and treatment. All in all, this research further investigates the molecular pathogenesis of LUAD, which can be served to instruct in-depth research of LUAD. | true | true | true |
|
PMC9600308 | Defang Zhou,Longying Ding,Menglu Xu,Xiaoyao Liu,Jingwen Xue,Xinyue Zhang,Xusheng Du,Jing Zhou,Xiyao Cui,Ziqiang Cheng | Musashi-1 and miR-147 Precursor Interaction Mediates Synergistic Oncogenicity Induced by Co-Infection of Two Avian Retroviruses | 21-10-2022 | ALV-J,REV,Musashi-1,miR-147 precursor,synergistic tumorigenesis,co-infection,NF-κB/KIAA1199/EGFR pathway | Synergism between avian leukosis virus subgroup J (ALV-J) and reticuloendotheliosis virus (REV) has been reported frequently in co-infected chicken flocks. Although significant progress has been made in understanding the tumorigenesis mechanisms of ALV and REV, how these two simple oncogenic retroviruses induce synergistic oncogenicity remains unclear. In this study, we found that ALV-J and REV synergistically promoted mutual replication, suppressed cellular senescence, and activated epithelial-mesenchymal transition (EMT) in vitro. Mechanistically, structural proteins from ALV-J and REV synergistically activated the expression of Musashi-1(MSI1), which directly targeted pri-miR-147 through its RNA binding site. This inhibited the maturation of miR-147, which relieved the inhibition of NF-κB/KIAA1199/EGFR signaling, thereby suppressing cellular senescence and activating EMT. We revealed a synergistic oncogenicity mechanism induced by ALV-J and REV in vitro. The elucidation of the synergistic oncogenicity of these two simple retroviruses could help in understanding the mechanism of tumorigenesis in ALV-J and REV co-infection and help identify promising molecular targets and key obstacles for the joint control of ALV-J and REV and the development of clinical technologies. | Musashi-1 and miR-147 Precursor Interaction Mediates Synergistic Oncogenicity Induced by Co-Infection of Two Avian Retroviruses
Synergism between avian leukosis virus subgroup J (ALV-J) and reticuloendotheliosis virus (REV) has been reported frequently in co-infected chicken flocks. Although significant progress has been made in understanding the tumorigenesis mechanisms of ALV and REV, how these two simple oncogenic retroviruses induce synergistic oncogenicity remains unclear. In this study, we found that ALV-J and REV synergistically promoted mutual replication, suppressed cellular senescence, and activated epithelial-mesenchymal transition (EMT) in vitro. Mechanistically, structural proteins from ALV-J and REV synergistically activated the expression of Musashi-1(MSI1), which directly targeted pri-miR-147 through its RNA binding site. This inhibited the maturation of miR-147, which relieved the inhibition of NF-κB/KIAA1199/EGFR signaling, thereby suppressing cellular senescence and activating EMT. We revealed a synergistic oncogenicity mechanism induced by ALV-J and REV in vitro. The elucidation of the synergistic oncogenicity of these two simple retroviruses could help in understanding the mechanism of tumorigenesis in ALV-J and REV co-infection and help identify promising molecular targets and key obstacles for the joint control of ALV-J and REV and the development of clinical technologies.
Avian leukosis virus subgroup J (ALV-J), the sixth subgroup of the Alpharetrovirus genus in the family Retroviridae, was identified in meat-type breeder chickens in 1991 [1,2]. Reticuloendotheliosis virus (REV), a gammaretrovirus of the family Retroviridae, was identified in adult turkeys in 1958 [3,4]. ALV-J and REV are transmitted both horizontally and vertically [5,6], and are the most common naturally occurring simple retroviruses associated with neoplastic and immunosuppressive diseases in poultry [7,8,9,10]. Owing to the similar characteristics and widespread nature of ALV-J and REV in the field, co-infection is common [11,12,13,14,15,16] and is an important emerging problem in chicken flocks. Synergism in ALV-J and REV co-infected chicken flocks causes higher mortality, more serious growth retardation, and immunosuppression [17,18], and facilitates viral replication and exosomal miRNA accumulation [19]. Synergism is a common phenomenon in retroviruses [20,21,22,23]. ALV-J induces late-onset myelocytomas, hemangiomas, and various other tumors in chickens, and REV induces chronic lymphomas in chickens, ducks, geese, pheasants, quails, and turkeys [24]. The two avian retroviruses continue to be of great interest in understanding the molecular mechanisms of tumorigenesis. In ALV-J and REV, the induction of neoplasms occurs in a minority of cases and only after several months of infection, presenting great difficulties in exploring the mechanisms of synergistic tumorigenesis. Since the suppression of cellular senescence and the activation of EMT are prerequisites for neoplasm and metastasis, many studies have considered cellular senescence and EMT as essential indicators of cellular oncogenicity in vitro [25,26,27,28,29,30]. Co-infection with oncogenic viruses has been increasingly reported in recent years [31,32]. Co-infection plays an important role in the development of neoplasms. The two co-infected oncogenic viruses can assist each other during the initiative and developmental process of the tumor. For retrovirus, slow (cis-activation) and acute (trans-activation) transformation are two classical oncogenicity mechanisms [6]. Slowly transforming retroviruses are considered replication-competent and do not carry oncogenes [33] that induce late-onset tumors through insertional mutagenesis by activating cellular proto-oncogenes or inactivating tumor suppressor genes [34,35,36,37,38]. If a slowly transforming retrovirus obtains a proto-oncogene from the host genome, it becomes an acutely transforming virus that can induce rapid-onset tumors within a short time after host infection, such as myc in MC29, CMII and OK10 strains of ALV [6], erbB in avian erythroblastosis virus [39], and v-rel in the T-strain of REV [40]. However, it has not been fully elucidated whether the synergism of the two oncogenic viruses further affects the interaction between the viruses and host cells or can further modify signaling pathways. In the present study, we revealed a synergistic oncogenicity mechanism induced by two simple oncogenic retroviruses, ALV-J and REV, which can help to identify promising molecular targets and key obstacles for the joint control of ALV-J and REV and in the development of clinical technology.
Primary chicken embryo fibroblasts (CEFs), DF-1 cells (a spontaneously immortalized CEF cell line), and human 293T cells, maintained in the laboratory of animal pathology of Shandong Agriculture University, were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, and 1% l-glutamine, and incubated at 37 °C in a 5% CO2 incubator. The stock SNV strain of REV at 103.2 50% tissue culture infectious dose (TCID50) and the NX0101 strain of ALV-J at 103.8 TCID50 were maintained in the laboratory of animal pathology of Shandong Agriculture University. The TCID50 of the SNV and NX0101 strains was titrated by limiting dilution in the DF-1 culture. KIAA1199 and NF-κB p50 3′UTRs were cloned downstream of the luciferase reporter gene of the pmirGLO control vector to create wild-type pmirGLO-KIAA1199 3′UTR (WT KIAA1199) and pmirGLO-p50 3′UTR (WT p50) plasmids, respectively (GenePharma, Shanghai, China). Bioinformatics analysis software tools and websites, including RNA22, NCBI, ENSEMBLE, and the RNA-protein interaction prediction website (http://pridb.gdcb.iastate.edu/RPISeq/, accessed on 28 September 2022), were used to analyze and predict the binding sites between miR-147 and the 3′ UTR regions of KIAA1199 or NF-κB p50. The pmirGLO-KIAA1199 3′ UTR mutant plasmid (Mut KIAA1199) and the pmirGLO-p50 3′ UTR mutant plasmid (Mut p50) were constructed through site-directed mutagenesis. miR-147 mimics, miR-147 inhibitors, Flag-MSI1, MSI1 Cas9/gRNA, Flag-KIAA1199, KIAA1199 Cas9/gRNA, KIAA1199 shRNAs, Flag-NF-κB p50, NF-κB p50 Cas9/gRNA, NF-κB p50 shRNAs, Flag-ALV-J gag, Flag-ALV-J pol, Flag-ALV-J env, HA-REV gag, HA-REV pol, and HA-REV env plasmids were purchased from GenePharma (Shanghai, China). The construction of MSI1 mutant plasmids was performed using the Fast Site-Directed Mutagenesis Kit (TIANGEN, Beijing, China) according to the manufacturer’s instructions.
This study was performed in strict accordance with the recommendations of the Shandong Institutional Animal Care and Use Committee. Ethical approval for this study was obtained from the Ethics Committee of Animal Experiments in Shandong Province (permit no. SDAU 15-124). Specific pathogen-free chicken embryos (120 embryos) were purchased from SPAFAS Co. (Jinan, China; a joint venture with Charles River Laboratory, Wilmington, MA, USA), allocated into four groups, and placed in separate incubators supplied with filtered positive-pressure air. Thirty embryos were each inoculated with ALV-J (100 μL, 103.8 TCID50) and 100 μL DMEM per egg, or REV (100 μL, 103.2 TCID50) and 100 μL DMEM per egg, or both ALV-J (100 μL, 103.8 TCID50) and REV (100 μL, 103.2 TCID50) per egg, through the allantoic cavity at an embryonic age of 6 days. Mock-infected embryos were inoculated with 200 μL DMEM. Of these embryos, 15, 19, and 10 hatched in the ALV-J (15/30; 50.0%), REV (19/30; 63.3%), and co-infection groups (10/30; 33.3%), respectively. In comparison, 90% (27/30) of the uninoculated control chickens hatched. The chickens were observed daily and euthanized when they were apparently ill or at 18 weeks of age. Of the 44 virus-infected chickens, 13 died at weeks 3, 4, and 10 for reasons unrelated to the infection. Ten chickens in the ALV-J group, 13 chickens in the REV group, and 8 chickens in the co-infection group were euthanized at 18 weeks of age, and 4 of 10, 0 of 13, and 7 of 8 bore tumors. In total, tumor tissues were obtained from 4 ALV-infected and 7 co-infected chickens. Myelocytomas, fibromas, lymphomas, and corresponding non-cancerous tissues, including the bone marrow, liver, and heart from each chicken, were divided into three portions for Western blotting (WB), ELISA, and RNA extraction assays.
The cancerous tissues and corresponding non-cancerous tissues of chickens were formalin-fixed, paraffin-embedded, sectioned, and stained for histopathological observation.
Total RNA from infected CEF cell samples, either mock-infected or infected with ALV-J or REV alone, or co-infected with both ALV-J and REV for 72 hpi, was separated on 15% agarose gels to extract small RNA (18–30 nt). Illumina small-RNA deep sequencing was performed as previously described [19].
SDT buffer was added to the CEF cell samples, and the same batch of samples was used for Illumina small RNA deep sequencing. The lysate was sonicated and boiled for 15 min. After centrifugation at 14,000× g for 40 min, the supernatant was quantified using a BCA Protein Assay Kit (Bio-Rad, Hercules, CA, USA). Proteins from each sample were separated using SDS-PAGE, prepared using a sample filter-aid, and labeled according to the manufacturer’s instructions (Applied Biosystems, Carlsbad, CA, USA). The TMT-labeled peptides were fractionated by SCX chromatography using an AKTA Purifier system (GE Healthcare, Chicago, IL, USA). Each fraction was subjected to nanoLC-MS/MS analysis. LC-MS/MS analysis was performed using a Q Exactive mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). MS/MS spectra were searched using the MASCOT engine (version 2.2; Matrix Science, London, UK) embedded in the Proteome Discoverer 1.4.
The miRNA mimic, KIAA1199 3 UTR, or NF-κB p50 3′ UTR luciferase reporter plasmids were co-transfected into CEF cells. At 48 hpi, cell lysates were prepared according to the manufacturer’s instructions using the Dual-Lumi™ Luciferase Reporter Gene Assay Kit (Beyotime Co., Ltd., Shanghai, China). Luciferase activity was measured using the dual-luciferase reporter assay system (Beyotime Co., Ltd.) and normalized against the activity of the Renilla luciferase gene.
The cells were lysed using cell lysis buffer (Beyotime) and incubated on ice for 5 min. Lysates were resuspended in SDS loading buffer, boiled for 5 min, loaded, and run on a 10% SDS-PAGE gel, and then transferred onto a nitrocellulose membrane (Solarbio, Beijing, China). Membranes were blocked with 5% skimmed milk at 4 °C overnight and probed with anti-ALV-J env (mouse monoclonal W459, Animal Pathology Lab, Shandong Agriculture University, Taian, China), anti-REV env (mouse monoclonal W460, Animal Pathology Lab, Shandong Agriculture University), anti-N-cadherin (rabbit monoclonal ab76011, Abcam, Cambridge, UK), anti-E-cadherin (rabbit monoclonal ab40772, Abcam), anti-vimentin (rabbit monoclonal ab92547, Abcam), anti-SNAIL (rabbit polyclonal ab85936, Abcam), anti-Flag (mouse monoclonal AT0022, Engibody, DE, USA), anti-HA (mouse monoclonal AT0024, Engibody), anti-KIAA1199 (rabbit polyclonal bs-21528R, Bioss, Beijing, China), anti-EGFR (rabbit polyclonal bs-10007R, Bioss), anti-MSI1 (rabbit monoclonal ab52865, Abcam), and anti-p50 (rabbit polyclonal bs-1194R, Bioss) antibodies at 1:1000, 1:1000, 1:1000, 1:1000, 1:1000, 1:1000, 1:1000, 1:1000, 1:200, 1:400, 1:2000, and 1:1000 dilutions, respectively, followed by horseradish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibody (Engibody) or HRP-conjugated goat anti-mouse secondary antibody (Engibody) at a dilution of 1:3000. β-actin was used as the loading control. Protein levels were detected using the Enhanced HRP-DAB Chromogenic Substrate Kit (Tiangen), according to the manufacturer’s instructions.
The specific primer sequences for pri-miR-147, KIAA1199, NF-κB p65, NF-κB p50, MSI1, EGFR, and GAPDH used in this study are listed in Table S1. Total RNA from CEF cells that had been either mock-infected, mono-infected with ALV-J or REV, or co-infected with ALV-J and REV was isolated using the Tiangen RNeasy mini kit (TIANGEN) according to the manufacturer’s instructions, with optional on-column DNase digestion. RNA integrity and concentrations were assessed by means of agarose gel electrophoresis and spectrophotometry, respectively. RNA (1 µg per triplicate reaction) was reverse-transcribed to cDNA using the Taqman Gold Reverse Transcription kit (Applied Biosystems). Real-time RT-PCR (qRT-PCR) was performed using SYBR® Premix Ex Taq, and specific primers (Table S1). All values were normalized to endogenous GAPDH levels to control for variation. For qRT-PCR analysis of miR-147, we used an miRcute miRNA first-stand cDNA synthesis kit and an miRcute miRNA qPCR detection kit (SYBR Green) (TIANGEN). The reverse primer provided in the miRcute miRNA qPCR detection kit was complementary to the poly (T) adapter. Data were collected on an ABI PRISM 7500 and analyzed using Sequence Detector v1.1 software (Applied Biosystems, USA). All values were normalized to endogenous U6 to control for variations. The primers specific for U6 are listed in Table S1. Assays were performed in triplicate, after which the average threshold cycle (CT) values were used to determine relative concentration differences based on the ΔΔCT method of relative quantization described in the manufacturer’s protocol.
The translocation of NF-κB p65 from the cytoplasm to the nucleus was examined via immunofluorescence. CEFs were washed with PBS and fixed with 4% paraformaldehyde for 20 min. After fixation, the cells were permeabilized with 0.25% Triton X-100 in PBS for 10 min and blocked with 10% BSA in PBS for 1 h. Anti-NF-κB p65 antibody (1:100, Bioss) was incubated overnight at 4 °C, followed the next day by a one-hour incubation at room temperature with anti-rabbit IgG antibody labeled with FITC (1:1000, Engibody). Finally, the cells were washed with PBS, incubated with DAPI for 5 min, and then observed under a fluorescence microscope after washing with PBS.
A senescence β-galactosidase staining kit was purchased from Beyotime (Shanghai, China) and was used to evaluate cellular senescence according to the manufacturer’s instructions.
Chicken NF-κB p65 and chicken NF-κB P-IκBα ELISA kits were purchased from Senbeijia (Nanjing, China) and were used to determine the expression levels of NF-κB p65 according to the manufacturer’s instructions.
The RNA ChIP kit was purchased from Active Motif (Shanghai, China) and was used to assay RNA–protein interactions according to the manufacturer’s instructions.
Data are presented as the mean ± standard deviation(s). The t-test and one-way ANOVA tests were performed using SPSS v. 13.0 statistical software (SPSS, Chicago, IL, USA). Statistical significance was set at p ≤ 0.05.
The suppression of cellular senescence or activation of EMT is a prerequisite for tumorigenesis and metastasis. To understand whether ALV-J and REV synergistically affected oncogenicity, cellular senescence and EMT were measured in ALV-J and REV co-infected cells. The RNA levels and protein levels of both ALV-J and REV were increased significantly in co-infected cells compared to those in mono-infected cells (Figure 1A,B,E–G), confirming that there was synergistic replication between ALV-J and REV. Senescence-associated (SA)-β-Gal staining revealed that ALV-J and REV synergistically inhibited cellular senescence (Figure 1C,D). The expression levels of EMT-associated proteins, assessed via Western blotting (WB) analysis, suggested that ALV-J and REV synergistically activated the EMT process (Figure 1E–G). These findings confirm that ALV-J and REV synergistically suppress cellular senescence and activate the EMT process in vitro.
Chick embryo fibroblasts (CEFs) co-infected with ALV-J and REV, mono-infected with ALV-J or REV, and mocks were analyzed using tandem mass tag (TMT)-based proteomics combined with miRNA whole-genome sequencing analysis. Among the 33 differentially expressed proteins and 17 differentially expressed miRNAs (Figure 2A,B), only miR-147 (or pri-miR-147) exhibited potential interactions with Musashi-1 (MSI1), NF-κB p50, and KIAA1199, which are associated with the cancer signaling pathway. The decreased miR-147 expression was verified using qPCR (Figure 2C), and the increased expression of MSI1, KIAA1199, and NF-κB p50 were verified using WB in co-infected CEFs (Figure 2D). RNA–protein interaction analysis showed interactions between miR-147 and MSI1, KIAA1199, and NF-κB p50 in ALV-J and REV co-infected cells. These results suggest that the interaction between miR-147 and activated MSI1, KIAA1199, and NF-κB p50 may be important in the synergistic oncogenicity induced by ALV-J and REV.
To verify the association of miR-147, MSI1, KIAA1199, and NF-κB p50 with oncogenicity, we measured cellular senescence and EMT through the construction and transfection of miR-147 mimics, miR-147 inhibitors, FLAG-MSI1, MSI1 Cas9/gRNA, FLAG-KIAA1199, KIAA1199 Cas9/gRNA, KIAA1199 shRNAs, FLAG-NF-κB p50, NF-κB p50 Cas9/gRNA, and NF-κB p50 shRNAs in CEF cells. SA-β-Gal staining and WB assays revealed that the miR-147 inhibitor MSI1, KIAA1199, or NF-κB p50 blocked cellular senescence and promoted EMT (Figure 3A–D), whereas miR-147 mimics, MSI1 knockdown, KIAA1199 knockdown, and NF-κB p50 knockdown promoted cellular senescence and inhibited EMT (Figure 3E–H). These data suggest that miR-147, MSI1, KIAA1199, and NF-κB p50 are associated with oncogenicity.
We intended to detect the expression of the miR-147 precursor (pri-miR-147) to determine whether mature miR-147 was inhibited before or after pri-miR-147 transcription. In contrast to mature miR-147, ALV-J and REV synergistically activated the expression of pri-miR-147 rather than inhibiting it. This showed that miR-147 was inhibited after pri-mir147 transcription (Figure 4A). The relationship between MSI1 and pri-miR-147/miR-147 was further analyzed. MiR-147 levels were detected when MSI1 was over- or under-expressed in CEFs. When MSI1 was overexpressed, the levels of mature miR-147 were suppressed more than 3.77-fold (Figure 4B). Upon MSI1 knockdown, the levels of mature miR-147 were elevated more than 5.58-fold (Figure 4C). Furthermore, RNA chromatin immunoprecipitation (ChIP) revealed over 15.6-fold enrichment of pri-miR-147, which is associated with MSI1 (Figure 4D,E), indicating that MSI1 directly targeted pri-miR-147 RNA. To identify the domain in MSI1 that binds pri-miR-147, we constructed four MSI1 mutants based on its RNA-binding sites [41,42], and transfected CEFs to detect mature miR-147 levels. The four MSI1 mutants are shown in Figure 4F. WB analysis confirmed that all MSI mutants were successfully transfected into the CEF cell line DF-1 (Figure 4G). The miR-147 expression level showed that only MSI1 mut1 relieved the inhibition of miR-147 maturation (Figure 4H), implying that the RNA-binding site (amino acid sequences 33 and 35 to 39) was the key domain for inhibiting miR-147 maturation. These data suggest that MSI1 directly targets pri-miR-147 through its RNA binding site (amino acid sequences 33 and 35 to 39), leading to the inhibition of miR-147 maturation.
To confirm that miR-147 directly targeted KIAA1199 and NF-κB p50, a dual-luciferase assay was performed in CEFs. The KIAA1199 3′ untranslated region (UTR) luciferase reporter assay revealed that miR-147 significantly inhibited the activity of the KIAA1199 3′ UTR reporter and that of the NF-κB p50 3′ UTR reporter, but not that of control reporters (Figure 5A,B). miR-147 inhibited the activities of the KIAA1199 3′ UTR reporter and the NF-κB p50 3′ UTR reporter and suppressed the endogenous expression levels of KIAA1199 and NF-κB p50 in CEF cells in a dose-dependent manner (Figure 5C–E). In contrast, the miR-147 inhibitor upregulated the expression of endogenous KIAA1199 and NF-κB p50 in a dose-dependent manner (Figure 5F). Bioinformatics analysis identified one putative miR-147 binding site at the KIAA1199 3′ UTR and the NF-κB p50 3ʹ UTR, respectively (Figure 4G,H). Mutations in the putative miR-147 binding site eliminated the inhibitory effect of miR-147 on the reporter activities of the KIAA1199 3′ UTR and NF-κB p50 3′ UTR (Figure 5I,J). These findings suggested that miR-147 directly targeted KIAA1199 and NF-κB p50.
Previous studies have shown that KIAA1199, an oncogene that is transcriptionally induced by NF-κB proteins, promotes EGFR stability and contributes to the activation of NF-κB/EGFR signaling pathway crosstalk in breast cancer [43,44]. To determine whether EGFR is involved in NF-κB/KIAA1199 signaling that is synergistically activated by ALV-J and REV, expression levels of NF-κB p65, phosphorylated IκBα, and EGFR were detected using qPCR, ELISA, and WB. Compared to mono-infection, NF-κB p65 RNA levels in co-infected cells were elevated 1.78- and 1.91-fold, respectively (Figure 6A). ELISA was used to confirm these results (Figure 6B). Compared with single infection, translocation of NF-κB to the nucleus was significantly observed in CEF-co-infected ALV-J and REV (Figure S1). The increase in phosphorylated IκBα levels indicated that ALV-J and REV synergistically activated the NF-κB signaling pathway (Figure 6C). Furthermore, we found that ALV-J and REV synergistically enhanced EGFR expression levels (Figure 6D,E). To determine the correlation between NF-κB, KIAA1199, and EGFR in ALV-J and REV co-infected cells, RNA interference was carried out via the construction and transfection of NF-κB p65 or KIAA1199 shRNAs into DF-1 cells. Upon NF-κB p65 or KIAA1199 knockdown, the expression levels of KIAA1199 and EGFR (Figure 6F–H) or EGFR and NF-κB p65 were suppressed (Figure 6I–K). Taken together, these findings suggested that ALV-J and REV synergistically activate the NF-κB/KIAA1199/EGFR signaling pathway.
The genome of simple retroviruses is composed of gag, pol, and env, which encode core proteins, proteases, and envelope proteins, respectively [45]. To determine whether the synergistic activation of MSI1 by ALV-J and REV is caused by its structural proteins, we detected the expression of MSI1 in cells infected with REV and transfected with ALV-J structural proteins (gag, pol, and env), or infected with ALV-J and transfected with REV structural proteins (gag, pol, and env). All viral structural proteins showed the synergistic promotion of MSI1 expression; however, ALV-J gag and REV gag showed the most significant promotion effect on MSI1 expression (Figure 7A,B). Furthermore, co-transfection of the two gags from ALV-J and REV demonstrated synergism in regard to MSI1 activation (Figure 7C). These findings confirmed that all structural proteins of ALV-J and REV were involved in the synergistic activation of MSI1.
To verify the synergistic oncogenicity occurring in ALV-J and REV co-infected tumor-bearing chickens, we established a tumor model induced by ALV-J and REV and measured the key molecules and the NF-κB/KIAA1199/EGFR pathway in three type tumors. All chickens were euthanized at 18 weeks of age; 4 of 10 (40%) ALV-J infected chickens, 0 of 13 (0%) REV-infected chickens, and 7/8 (87.5%) co-infected chickens bore tumors. Histopathological examination showed that the tumors induced by ALV-J were myelocytomas; however, ALV-J and REV induced myelocytomas, lymphomas, and endocardial fibromas (Figure 8A). These findings suggested that ALV-J and REV synergistically promoted tumorigenesis in chickens. To validate the results of the in vitro experiments, we detected the RNA levels of miR-147, MSI1, KIAA1199, NF-κB p50, and EGFR in the cancerous tissues of chickens in different infection groups, including the bone marrow, liver, and heart. Compared to the mono-infection group, the RNA levels of miR-147 were significantly downregulated, whereas MSI1, KIAA1199, NF-κB p50, and EGFR were significantly upregulated in the bone marrow, livers, and hearts of chickens (Figure 8B–F). Next, we determined the RNA expression levels of miR-147 in five cases of myelocytomas, lymphomas, and endocardial fibromas. Compared to non-cancerous tissues in co-infected chickens, miR-147 levels in myelocytomas, fibromas, and lymphomas in four, five, and five cases, respectively, were suppressed (Figure 9A). Compared to non-cancerous tissues, ELISA revealed that four, five, and four cases of phosphorylated IκBα levels were elevated in myelocytomas, fibromas, and lymphomas, respectively (Figure 9B). Simultaneously, elevated expression levels of EGFR, KIAA1199, NF-κB p50, and MSI1 were also confirmed in myelocytomas, fibromas, and lymphomas using WB (Figure 9C–E). Compared to the corresponding non-cancerous tissues, MSI1 and NF-κB/KIAA1199/EGFR pathway crosstalk was upregulated in 15 of 15 and 13 of 15 tumors, respectively, whereas miR-147 was downregulated in 14 of 15 tumors (Figure 9F). These data suggest that a synergistic tumorigenesis mechanism occurred in chickens co-infected with ALV-J and REV.
Synergistic interactions between two retroviruses in co-infected hosts have been well documented [20,21,22,23]. Co-infection with two or more oncogenic retroviruses is known to accelerate cancer development [32]. Recent studies have shown that co-infection of ALV-J and REV causes higher mortality, more serious growth retardation, and immunosuppression, facilitating viral replication and changing the miRNA expression profile [17,18,19]. However, the question of whether synergism promotes oncogenicity and the underlying synergistic mechanism remains unclear. In this study, we found that ALV-J and REV synergistically suppressed cellular senescence and activated EMT in vitro, indicating that these two viruses have developed strategies to synergistically promote oncogenic potential in vitro. The suppression of cellular senescence and the activation of EMT are prerequisites for neoplasm and metastasis, which have been commonly considered as essential indicators of cellular oncogenic potential in vitro [25,26,27,28,29,30]. To identify the key molecules responsible for the synergistic oncogenicity induced by ALV-J and REV in host cells, TMT-based proteomics, combined with miRNA whole-genome sequencing, was used to screen and identify the key molecules in the co-infected/mono-infected/mock cells. Interestingly, an miRNA molecule, miR-147, known as a tumor suppressor [46,47,48,49] showed ectopic expression, which increased in mono-infected cells and decreased in co-infected cells. However, its precursor, pri-miR-147, showed a synergistic increase in ALV-J and REV-co-infected cells. These data suggested that a certain molecule blocks miR-147 maturation, releasing its target signals for tumorigenesis. Thus, target analysis indicated that miR-147 or pri-miR-147 exhibited potential interactions with MSI1, NF-κB p50, and KIAA1199, which are associated with the cancer signaling pathway. Next, we demonstrated that miR-147, MSI1, NF-κB p50, and KIAA1199 play a critical role in cellular senescence suppression and EMT activation, indicating that these molecules are involved in the synergistic oncogenicity induced by ALV-J and REV. MSI1, an RNA-binding protein, has been found to regulate multiple critical biological processes that are relevant to cancer initiation and progression [50]. KIAA1199, a novel proto-oncogene, has been associated with tumor progression and metastasis in numerous cancers [51]. NF-κB activation is key to the early development of some cancers [52]. Based on the functions of these molecules, we speculated that MSI1 blocks the maturation of miR-147, which relieves the inhibition of NF-κB p50 and KIAA1199. The experimental results support our speculation. MSI1 directly targeted pri-miR-147 through its RNA-binding site, inhibiting miR-147 maturation. Because miR-147 directly targets NF-κB p50 and KIAA1199, downregulation of miR-147 promoted the upregulation of NF-κB p50 and KIAA1199. Recent studies have demonstrated that NF-κB and EGFR are partners in cancer, and NF-κB-induced KIAA1199 promotes EGFR stability, contributing to the activation of the NF-κB/EGFR signaling pathway [43,44]. Because EGFR was absent in the TMT-based proteomics results, we wanted to know whether EGFR is involved in the NF-κB/KIAA1199 signaling that is synergistically activated by ALV-J and REV. The results showed that EGFR participated in the NF-κB/KIAA1199 pathway, namely, NF-κB/KIAA1199/EGFR, which was synergistically activated by ALV-J and REV. We observed that ALV-J and REV synergistically activate MSI1, which binds pri-miR-147, blocking miR-147 maturation and thereby relieving the inhibition of the NF-κB/KIAA1199/EGFR signaling pathway. We then investigated how ALV-J and REV synergistically activate MSI1. The results showed that all structural proteins, especially gags from ALV-J and REV, synergistically activated the expression of MSI1. Generally, gag proteins from complex retroviruses or acutely transforming retroviruses (carrying gag-onc fusion genes) are involved in oncogenesis [53,54]. Here, we observed for the first time that gags from two simple retroviruses synergistically promoted oncogenicity. Finally, the key molecules and signaling pathways involved in synergistic tumorigenesis were verified in tumor-bearing chickens infected with ALV-J and REV.
In conclusion, the current study revealed a synergistic oncogenicity mechanism induced by two simple retroviruses, ALV-J and REV. ALV-J and REV synergistically suppressed cellular senescence and activated EMT in vitro, and synergistically induced tumorigenesis and tumor spectrum extension in vivo. Mechanistically, as shown in Figure 10, after co-infection with ALV-J and REV, the released or expressed structural proteins from ALV-J and REV in the cytoplasm synergistically activated MSI1 expression, which directly targeted pri-miR-147 through its RNA binding site, causing the inhibition of miR-147 maturation. This relieved the inhibition of the NF-κB/KIAA1199/EGFR signaling pathway, thereby suppressing cellular senescence and activating EMT. The synergistic oncogenicity mechanism of ALV-J and REV sheds light on the identification of promising molecular targets and key barriers to the joint control of ALV-J and REV and the development of clinical technologies. | true | true | true |
|
PMC9600392 | Stefania Evangelisti,Laura Ludovica Gramegna,Silvia De Pasqua,Magali Jane Rochat,Luca Morandi,Micaela Mitolo,Claudio Bianchini,Gianfranco Vornetti,Claudia Testa,Patrizia Avoni,Rocco Liguori,Raffaele Lodi,Caterina Tonon | In Vivo Parieto-Occipital White Matter Metabolism Is Correlated with Visuospatial Deficits in Adult DM1 Patients | 24-09-2022 | MR spectroscopy,metabolism,myotonic dystrophy,white matter,visuospatial functions | Myotonic dystrophy type 1 (DM1) is a genetic disorder caused by a (CTG) expansion in the DM protein kinase (DMPK) gene, representing the most common adult muscular dystrophy, characterized by a multisystem involvement with predominantly skeletal muscle and brain affection. Neuroimaging studies showed widespread white matter changes and brain atrophy in DM1, but only a few studies investigated the role of white matter metabolism in the pathophysiology of central nervous system impairment. We aim to reveal the relationship between the metabolic profile of parieto-occipital white matter (POWM) as evaluated with proton MR spectroscopy technique, with the visuoperceptual and visuoconstructional dysfunctions in DM1 patients. MR spectroscopy (3 Tesla) and neuropsychological evaluations were performed in 34 DM1 patients (19 F, age: 46.4 ± 12.1 years, disease duration: 18.7 ± 11.6 years). The content of neuro-axonal marker N-acetyl-aspartate, both relative to Creatine (NAA/Cr) and to myo-Inositol (NAA/mI) resulted significantly lower in DM1 patients compared to HC (p-values < 0.0001). NAA/Cr and NAA/mI correlated with the copy of the Rey-Osterrieth complex figure (r = 0.366, p = 0.033; r = 0.401, p = 0.019, respectively) and with Street’s completion tests scores (r = 0.409, p = 0.016; r = 0.341, p = 0.048 respectively). The proportion of white matter hyperintensities within the MR spectroscopy voxel did not correlate with the metabolite content. In this study, POWM metabolic alterations in DM1 patients were not associated with the white matter morphological changes and correlated with specific neuropsychological deficits. | In Vivo Parieto-Occipital White Matter Metabolism Is Correlated with Visuospatial Deficits in Adult DM1 Patients
Myotonic dystrophy type 1 (DM1) is a genetic disorder caused by a (CTG) expansion in the DM protein kinase (DMPK) gene, representing the most common adult muscular dystrophy, characterized by a multisystem involvement with predominantly skeletal muscle and brain affection. Neuroimaging studies showed widespread white matter changes and brain atrophy in DM1, but only a few studies investigated the role of white matter metabolism in the pathophysiology of central nervous system impairment. We aim to reveal the relationship between the metabolic profile of parieto-occipital white matter (POWM) as evaluated with proton MR spectroscopy technique, with the visuoperceptual and visuoconstructional dysfunctions in DM1 patients. MR spectroscopy (3 Tesla) and neuropsychological evaluations were performed in 34 DM1 patients (19 F, age: 46.4 ± 12.1 years, disease duration: 18.7 ± 11.6 years). The content of neuro-axonal marker N-acetyl-aspartate, both relative to Creatine (NAA/Cr) and to myo-Inositol (NAA/mI) resulted significantly lower in DM1 patients compared to HC (p-values < 0.0001). NAA/Cr and NAA/mI correlated with the copy of the Rey-Osterrieth complex figure (r = 0.366, p = 0.033; r = 0.401, p = 0.019, respectively) and with Street’s completion tests scores (r = 0.409, p = 0.016; r = 0.341, p = 0.048 respectively). The proportion of white matter hyperintensities within the MR spectroscopy voxel did not correlate with the metabolite content. In this study, POWM metabolic alterations in DM1 patients were not associated with the white matter morphological changes and correlated with specific neuropsychological deficits.
Myotonic dystrophy type 1 (DM1) is a multisystem disease associated with an unstable expansion of nucleotide repeats (CTG [cytosine–thymine–guanine] triplets) in the genes coding for myotonic dystrophy protein kinase (DMPK) [1] that contribute to the accumulation of mutant RNA aggregates, with consequent mis-splicing of downstream effector genes that affect almost all cells and organs of the human body [2,3,4]. Therefore, the disease is characterized by multisystem involvement with predominantly muscle and brain affection. Brain involvement in DM1 includes the presence of hyperostosis and whole brain atrophy, mostly evaluated in group analysis using advanced brain MRI techniques [5], and subcortical atrophy in form of increased ventricular volume that rarely can result in a severe cerebral ventriculomegaly with a normal pressure hydrocephalus-like appearance on MR imaging [6]. White matter involvement in DM1 has now been irrefutably demonstrated by a large body of literature that evidenced the presence of T2/fluid-attenuated inversion recovery (FLAIR) hyperintensities preferentially located bilaterally in frontal, temporal (temporo-polar in one-third of the cases) [5], and parietal lobes at periventricular and subcortical locations [7]. Advances in imaging technology recently provided substantial evidence for a diffuse involvement of the normal appearing white matter (NAWM) in DM1 patients, as demonstrated mainly by diffusion tensor imaging (DTI) techniques with mean diffusivity (MD) values increase and fractional anisotropy (FA) values reduction [5]. Brain proton MR spectroscopy (1H-MRS) is a non-invasive technique that can provide an in vivo evaluation of the biochemical profile within specific volumes of interest (VOI). It has been shown that a reduction of N-acetyl-aspartate (NAA) level is related to neuronal and/or axonal damage or loss as in neurodegenerative disorders, and an increase in myo-Inositol (mI) is considered a marker of glial reaction. Other metabolites that are typically quantified are creatine (Cr) and choline-containing compounds (Cho), respectively, markers of cellular energy and of cellular proliferation, increased membrane turnover, or inflammation [8]. MRS allows us to detect brain metabolic abnormalities that can be present at early stages of the diseases, when the structural morphological alterations are not evident, and findings of morphological MR imaging are still absent or ambiguous. Few previous studies, summarized in Supplementary Table S1, described brain metabolic profile in DM1 patients through the 1H-MRS technique [9,10,11,12,13,14], with heterogeneous technical parameters, VOI selection in regions including both gray and white matter and patients’ populations. The most common finding from previous studies is a reduction of NAA both as relative and absolute concentrations. Several studies have demonstrated that DM1 patients show a selective impairment in cognitive functioning, particularly concerning attention, memory, executive, and visuospatial domains [2,15,16,17,18,19]. Visuospatial functioning refers to cognitive processes necessary to “identify, integrate, and analyze space and visual form, details, structure and spatial relations in more than one dimension” [20]. Visuospatial processing, therefore, results from the integration of three main processes: 1. perception of basic visual elements such as light, contrasts, and orientation, principally mediated by the occipital cortices; 2. visual construction, resulting from the progressive integration of visual percepts with input from the parietal, temporal, and frontal cortices, and 3. visual memory (recall/recognition of visual information and topographical memory) [21]. Visuospatial deficit also impacts visuoconstructional abilities, which require the coordination of fine motor skills with spatial abilities to copy well-planned and organized geometric figures [22]. Both visuospatial and visuo-constructional impairments are considered distinctive features of the DM1 cognitive profile [23,24,25] and have been proposed as prognostic factors of cognitive and structural brain progressive degeneration [19,26]. Moreover, subjects with DM1 frequently present psychiatric comorbidities embracing personality and/or mood disorders, and a characteristic emotional imbalance [27,28]. In spite of their cognitive/affective difficulties and progressive motor impairment, patients with DM1 also may present anosognosia (or “lack of insight”), a mental disorder that impairs the ability to understand and perceive accurately one’s own illness [15,17,28]. The lack of disease awareness may seriously hinder early diagnostic assessment as it leads to secondary misattribution of symptoms and low compliance to treatment [17]. To date, the inability to be aware of one’s own medical condition has been correlated in DM1 patients to white matter alterations detected by morphological MR imaging [17,29]. The aim of the current study was to depict the metabolism of parieto-occipital white matter through the 1H-MRS technique in DM1 patients, while investigating the role of white matter pathology in neuropsychological dysfunction.
We enrolled thirty-seven consecutive adult patients with a genetic diagnosis of myotonic dystrophy type 1 (DM1 or “Steinert’s disease”) (F/M 22/15, age: 46.8 ± 11.7 years). Patients were recruited at the Neuromuscular Centre of the UOC Neurological Clinic of the IRCCS Institute of Neurological Sciences, Bologna, IT, during the period from 1 Junaury 2019 to 31 December 2021. The study protocol was approved by the local Ethical Committee (n. 1088-2020-OSS-AUSLBO). CTG triplet expansion sizes were measured in all patients in genomic DNA extracted by peripheral blood leukocyte using Southern blot analysis. Depending on the number of repeat expansions patients were classified into three genetic classes: E1 (50-150 CTG repeats), E2 (150-1000 CTG repeats), and E3 (more than 1000 CTG repeats). Depending on the age of disease onset, patients were stratified into three groups: congenital/childhood onset (from birth to 10 years of age), juvenile/adult onset (from 11 to 40 years of age), and late onset (more than 40 years of age). All DM1 patients underwent standardized clinical, neuropsychological, and MR spectroscopy and imaging evaluations. A cohort of ten sex- and age-matched healthy control subjects (HC) was also recruited for the quantitative MR evaluation. They were selected from the database of the Neuroimaging Laboratory, designed to collect normative values of MR parameters for clinical and research purposes. Demographic features of patients and controls are reported in Table 1.
An extensive neuropsychological assessment was performed on all patients. However, in the present study, we focus specifically on general cognitive functioning in addition to visuospatial and visuoconstructional abilities. Thus, each participant underwent a neuropsychological examination to obtain a clinical profile that included a general cognitive screening test (Mini-Mental State Examination, MMSE [30]) and a test assessing culture-free logical reasoning (Raven’s Colored Progressive Matrices Test, CPM-47 [31]). Visuospatial abilities were investigated by The Benton Judgment of Line Orientation Test (BJLOT [32]) and the Street’s Completion Test [33]. The copy of the Rey-Osterrieth complex figure (ROCF-copy [34]) further explored visuoconstructional abilities. Patients’ raw scores were corrected according to Italian normative values [21,35]. Percentages of impairment of DM1 patients who showed significant neuropsychological dysfunctions across different cognitive domains were established using Italian normative data for both, score adjustment (sex, age, and education) and the definition of cut-off thresholds. Finally, levels of self-awareness (or the presence of anosognosia) were measured with the Measurement of Anosognosia Instrument [36]. This questionnaire consists of a series of dichotomous items assessing the patients’ performance in selected cognitive domains (i.e., attention, memory, language, executive functions) and in daily life settings and was administered with both self-report and informant-rating versions. The responses provided by the patient in the double modality were compared to quantify the number of discrepant answers. Discrepancy scores were used to quantify the presence of anosognosia [37,38].
Participants underwent a multimodal standardized brain MR protocol acquired with a 3T scanner (Siemens MAGNETOM Skyra) equipped with a high resolution 64-channel coil. The protocol included single-voxel proton MR spectroscopy (suppressed-waterPoint RESolved Spectroscopy, PRESS) technique within the parieto-occipital white matter (POWM, volume = 8 mL, echo time/repetition time TE/TR = 30/2000 ms, number of averaged fids = 64, duration ~2 min) [39], a volumetric T1-weighted sequence (3D MPRAGE, magnetization prepared rapid gradient-echo, sagittal acquisition, isotropic voxel 1 × 1 × 1 mm3, no slice gap, TE = 2.98 ms, TR = 2.300 ms, Inversion Time IT = 900 ms, flip angle = 9°, acquisition matrix = 256 × 256, pixel bandwidth = 240 Hz, GRAPPA acceleration factor = 2, duration ~5 min) and a volumetric fluid-attenuated inversion recovery (FLAIR) T2-weighted sequence (3D SPACE, sagittal acquisition, isotropic voxel 1 × 1 × 1 mm3, no slice gap, TE = 428 ms, TR = 5000 ms, IT = 1.800 ms, flip angle = 120°, acquisition matrix = 256 × 256, pixel bandwidth = 780 Hz, GRAPPA acceleration factor = 2, duration ~5 min).
All MRI images were analyzed by a neuroradiologist with 8 years of experience in reporting neurodegenerative disorders who evaluated the presence of structural brain MRI alterations in DM1 patients, including the presence of white matter changes appearing as hyperintensities on T2- weighted sequence and normal pressure hydrocephalus-like appearance. In particular, the severity of white matter changes appearing as hyperintensities on the T2- weighted sequence was estimated using a modification of the age-related white matter change rating scale (Fazekas scale) [40]. This scale is scored from 0 to 3 as follows: (0) Absent, absence of periventricular or subcortical lesions (a single punctate hyperintensity less than 5 mm in size within the subcortical and periventricular white matter was considered normal); (1) Slight, continuous periventricular lines less than 5 mm in length and/or various non-confluent subcortical white matter foci less than 5 mm in size; (2) Moderate: continuous periventricular lines 5 to 10 mm long and/or subcortical white matter foci 5 to 10 mm in size beginning to merge; and (3) severe: periventricular stripes more than 10 mm long and/or irregular confluent lesions more than 10 mm in size. Regarding the presence of the normal pressure hydrocephalus-like appearance, the z- EVANS index was calculated based on the ratio between the maximum z-axial length of the frontal horns and the maximum cranial z-axial length on the coronal plane, which was perpendicular to the anteroposterior commissure plane on the anterior commissure as previously described [41]. Similarly, the callosal angle, considered the angle of the roof of the lateral ventricles on the coronal plane at the posterior commissure level [42] was reported. The presence of hyperostosis, enlarged perivascular spaces, and incidental MRI findings were reported as well.
Patients’ white matter changes appearing as white matter hyperintensities on FLAIR T2-weighted images were quantified through Jim software (Version 7.0, Xinapse Systems, Northants, UK, http://www.xinapse.com (accessed on 22 July 2022)), a semi-automatic threshold-based method specific for white matter lesions load segmentation. FLAIR T2-weighted images were linearly (FLIRT, part of the Oxford FMRIB Software Library FSL) co-registered to the T1-weighted images and the transformation matrix was then used to align the lesion map to the T1-w image. In this way, the accuracy of brain tissue segmentation could be improved by reducing the intensity contrast in T1-w images within known WM lesions by using the FSL lesion filling tool [43,44]. T1-weighted images were also non-linearly (FNIRT) aligned to the MNI (Montreal Neurological Institute) template, and the transformation was applied to the lesion maps and a group map was built by summing individual maps together.
Assessment of spectra quality was based on visual inspection, evaluating the absence of artifacts and the baseline, along with the signal to noise ratio (SNR ≥ 14) and Full Width at Half Maximum (FWHM ≤ 3.7) evaluation. Examples of the measured spectra can be seen in Figure 2D. Spectra were analyzed with the fitting software LCModel, version 6.3 [45]. N-Acetyl-aspartate (NAA), choline (Cho)-containing compounds, and myo-Inositol (mI) content was evaluated relative to creatine (Cr) or mI as internal references. As a measure of post processing model fitting quality, LCModel estimated fitting error <15% was considered.
Estimates of the tissue fractions within the MRS voxels was performed for each subject by using FSL-MRS (Figure 2B) (in particular, Siemens DICOM spectral data were converted into NIfTI with spec2nii tool, then svs_segment was applied after that the lesion-filled T1-weighted image was segmented into grey matter (GM), white matter (WM) and cerebrospinal fluid (CSF), with the tool fast [46]. In this way, potential bias related to the different tissue-volume fractions can be excluded, being more methodologically accurate and of support for the interpretation [47]. Moreover, the amount of WM lesion volume that was included within the MRS voxel was also assessed (Figure 2C), in order to calculate the amount of normal-appearing vs. altered WM included in the metabolic profile evaluation.
The normality of data distributions was assessed using the Shapiro-Wilk test. When normal continuous variables were compared between two groups, t-tests were used (or ANCOVA when covariates of no interests and/or more than two groups were included in the model), otherwise, Wilcoxon tests were applied. In MRS data comparisons, age and sex were added as covariates of no interest. Categorical variables were compared with chi-square tests. Pearson’s or Spearman’s correlations were performed according to the distribution. The statistical analyses were conducted with R-software version 3.5.2 (https://www.r-project.org/ (accessed on 22 July 2022)). Data from patients belonging to the E3 genetic class are reported for information, but not included in group comparison analysis given that only two patients belonged to this class. In order to evaluate the accuracy of proton MR spectroscopy parameters in discriminating DM1 patients from HC, a receiver operating characteristics (ROC) curve analysis was performed. The optimal cutoff value corresponded to the higher Youden’s index.
Among the 37 recruited DM1 patients, one did not complete the MR protocol and two had suboptimal quality spectra (due to motion artifacts, related lipids contamination from the scalp, and imprecise VOI localization). Therefore, 34 good quality spectra were included in the analysis, along with all 10 spectra from HC. Among the patients included, four had congenital/childhood onset, 22 juvenile/adult onset, and eight late had onset. According to the CTG repeats, 13 of them belonged to E1, 19 to E2, and two to E3 classes. Main descriptive data for patients and HC are reported in Table 1.
Considering the whole cohort of DM1, the general cognitive profile assessed by MMSE was impaired in 11.8% (4/34) of them (mean corrected score ± sd 26.5 ± 2.7, range 18.7–30.0), mostly driven by E2 (15.8%; 26.6 ± 2.8 18.7–30.0) and E3 (50%; 21.9 ± 4.2; 18.9–24.9) pathological scores. The deficit in logical reasoning, namely the ability to reason organized spatial perceptions into systemically related whole (CPM-47) was found in 14.7% (26.4 ± 5.7; 10.8–36.0) of DM1 patients, affecting E2 (15.8%; 25.9 ± 5.3; 14.8–33.3) and all E3 patients (14.1 ± 4.7; 10.8–17.4), while E1 mean performance ranked within a normal range (28.8 ± 3.6; 23.7–36.0). Both visuospatial and visuoconstructional functions were remarkably impaired in all DM1 patients, with a performance following a worsening trend across the three classes of nucleotide triplet’s expansion. Specifically, inaccurate judgments of lines orientation (BJLOT) were assessed in 26.5% (9/34) (22.5 ± 7.2; 6.0–30.0) of the patients, that is 15.4% of E1 (24.9 ± 6.1; 10.0–30.0), 26.3% of E2 (22.5 ± 6.3; 8.0–30.0), and 100% E3 (7.5 ± 2.1;6.0–9.0). Moreover, 16/34 that is 47.1% (27.3 ± 7.8; 8.0–36.0) of them ((23.1% of E1 (29.8 ± 7.4; 8.0–36.0), 57.9 % of E2 (27.3 ± 6.2; 8.0–34.6), and 100% of E3 (10.8 ± 2.5; 9.1–12.5) failed to accurately copy a complex geometrical figure (ROCF-copy), which is suggestive of deficits in motor planning and visuoconstructional abilities. Finally, deficits in visual perceptual organization, as measured with SCT, were found in only 5/34 that is 5.9% of DM1 patients (6.4 ± 2.3; 1.3–10.2), pertaining to E1 (7.7%; 5.4 ± 2.2; 1.3–8.6) and E2 (5.3%; 7.4 ± 1.9; 3.1–10.3) classes. Finally, discrepancy scores between self- and informant-ratings of patient’s cognitive performance were observed in most of the DM1 cohort (20/34, 58.8%), where 17 out of 34 patients (50%) reported an overestimation of their deficit, while 3 out of 34 (8.8%; −1.3 ± 2.1; −5.0–3.0) reported a lack of awareness (anosognosia). The over- vs. underestimation ratio differed among DM1 classes: 7.7% of E1 (−1.0 ± 2.0; −4.0–3.0) patients presented anosognosia, while 53.9% of them overestimated their own deficits; the same trend was accentuated among E2 patients (10.5%, −1.3 ± 2.3; −5.0–2.0); underestimating patients versus 31.6% overestimating), while both E3 patients showed to overestimate their cognitive impairments. Neuropsychological findings are summarized in Table 2. When neuropsychological data were analyzed comparing patients pertaining to different age of onset groups, the main differences were observed between patients with a congenital onset and those with a juvenile one, essentially concerning worse general cognitive abilities (MMSE), logical reasoning (CPM-47) and visuospatial abilities (BJLOT) among the congenital onset DM1 patients (see Supplementary Table S2 for more details).
The extensive neuroradiological characterization based on conventional morphological MR images for the DM1 cohort is reported in Table 3. Overall, almost all DM1 patients (30/34, 88.2%) had a Fazekas-scale score greater than zero, and 18/34 (52.9%) presented white matter hyperintensities on the FLAIR T2- weighted sequence in the temporo-polar regions. Hyperostosis was present in 11/34 (32.4%) patients, mainly within the frontal regions, and only in four was diffuse. Cerebral ventriculomegaly was observed in the majority of patients (19/34, 55.9%) and was symmetrical in all except two. The z- EVANS index was on average 0.28 ± 0.04 and 6/34 (17.6%) patients also presented disproportionately enlarged subarachnoid-space hydrocephalus (DESH) [48]. On average the callosal angle was (112.8° ± 16.6°), and all the patients had an enlarged perivascular space within the hemispheric white matter. As for the whole brain white matter lesion load evaluation, it was performed in 33 over 34 patients since one (n.33) did not have good-quality FLAIR T2-weighted images due to movement artifacts. In the whole group of DM1 patients, the total lesion load was 7883 ± 12,347 mm3 (mean ± SD), 3732 [8474] mm3, (median [IQR]), 89–67,863 mm3 (range). When the three genetic classes were instead considered separately, it was (4301 ± 4574) mm3 (2995 [7085], 208–14,283 mm3) in E1 patients, (10184 ± 15,551) mm3 (4835 [10,232], 89–67,863 mm3) in E2 patients and 10,617 mm3 in the E3 patient. Even if a larger white matter lesion load seems to be present in E2 patients if compared to E1, this difference was not statistically significant (Wilcoxon test, p-value = 0.2231). To represent the overall spatial distribution of white matter changes in our DM1 cohort, the group lesion map is shown in Figure 1.
NAA/Cr and NAA/mI resulted in significantly lower in DM1 patients compared to HC (Table 4, Figure 2D,E). As for the genetic subgroups, no significant differences were found between E1 and E2 patients, but the ANCOVA highlighted a significant effect of group for NAA/Cr and NAA/mI due to the significant alterations in both E1 and E2 patients compared to HC (Table 4, Figure 2F). An analogous effect was observed when DM1 patients were stratified according to the age of onset: no significant differences were found between the subgroups of patients, but only compared to HC (Supplementary Table S3). The fraction of different tissues (WM, GM, and CSF) included within the MRS volume of interest (Figure 2B) did not differ significantly between groups (Table 5, Supplementary Table S4), excluding therefore a potential bias in the comparisons of metabolites content. As for the WM fraction included in the MRS VOI (Figure 2C), the ratios of normal-appearing WM and altered WM on FLAIR T2 weighted- images were instead different between the two classes E1 vs. E2 (Table 5). We therefore verified whether it was correlated with the metabolites content, and there were no correlations of the percentage of altered WM within the MRS VOI and NAA/Cr (Spearman rho = −0.064, p = 0.725), Cho/Cr (rho = 0.251, p = 0.158), mI/Cr (rho = 0.133, p = 0.461), NAA/mI (rho = −0.151, p = 0.401). These results suggest that the metabolic alterations evaluated with MRS are not related to the structural alteration of the tissue, but instead present also in normal-appearing WM.
NAA/Cr within POWM showed significant positive correlations with ROCF-copy (r = 0.366, p = 0.033) and performance Street’s completion test (r = 0.409, p = 0.016) scores. NAA/mI was also positively correlated with ROCF-copy (r = 0.401, p = 0.019) and with Street (r = 0.341, p = 0.048) tests scores. Moreover, both NAA/Cr and NAA/mI negatively correlated with disease duration (r = −0.530, and p = 0.001; r = −0.468, p = 0.005, respectively). The total WM lesion load was negatively correlated with the ROCF-copy test score (r = −0.360, p = 0.040,), whereas no significant correlations between metabolite content and total WM lesion load were found. Given the significant differences, NAA/Cr and NAA/mI were the selected parameter to evaluate 1H-MRS diagnostic accuracy. For NAA/Cr, the ROC curve analysis showed an area under the curve (AUC) of 0.94, with sensitivity and specificity respectively of 90% and 94% when considering a cut-off value of 2.101. As for NAA/mI, the AUC was 0.88, with sensitivity and specificity respectively of 100% and 65% when considering a cut-off value of 2.281. At the single-subject level, considering NAA/Cr, only two patients had unaltered values (Table 3).
Our study demonstrates that analysis of brain metabolism in the parieto-occipital white matter by proton magnetic resonance spectroscopy (1H-MRS) may assist in the neuroradiological diagnosis of DM1 by revealing pathological NAA/Cr ratio with high sensitivity and specificity. Moreover, this finding expanded previous studies performed by using other advanced MRI techniques such as tract-based spatial statistics (TBSS), and region of interest diffusion tensor imaging (ROI-based DTI) suggesting that alterations in white matter metabolism may be associated with deficits in visuospatial and visuoconstructional abilities, confirming involvement of the normal-appearing parieto-occipital white matter. Altered NAA/Cr as measured with 1H-MRS within the parieto-occipital white matter could identify DM1 patients with good accuracy, sensitivity, and specificity, although it was not significantly different between the genetic classes E1 and E2. Interestingly, for three patients this metabolic alteration was the only sign detected in the context of an apparently normal morphological brain MRI examination. This approach is based on a non-invasive and reproducible technique, with a relatively short acquisition time, that could be included in the clinical neuroradiological routine [8]. White matter alterations in DM1 are typically hyperintense changes in the FLAIR T2-weighted images. The pathophysiology of the WM involvement, as for the pathological alterations that underly the imaging phenotypes, are still under debate [2]. A recent review evaluated 41 pathological studies focusing on the microscopic brain alteration in a total of 130 DM1 patients [49]. Eight studies highlighted the presence of gliosis in the white matter in 37 DM1 patients overall, although the type of gliosis was mostly not specified (e.g., reactive astrogliosis, microgliosis/activated microglia cells, or oligodendrocyte response). The largest study among these eight included 11 patients and reported that a loss of axons accompanied the myelin loss [50], whereas in two case reports, myelin loss coincided with “relative axonal sparing” [51,52]. Moreover, Itoh et al., also found capillary hyalinization and fibrillary gliosis [50]. Interestingly, heterotopic neurons were found in white matter in three studies [53,54,55] that evaluated nine patients. Four brains pertained to patients with congenital DM1 that died shortly after birth; the other five were from DM1 patients. Heterotopic neurons were found both in brain subcortical and deep white matter. There are a few hypotheses on the pathogenetic mechanism of the observed histopathological alterations and, therefore, of the MRI signal alterations. Some studies have found the presence of abnormal mutant RNA accumulation in the nuclei of brain cells including oligodendrocytes [56,57]. It has been suggested that these foci of RNA accumulation are toxic for the cells as they compromise the regulation of alternative splicing in different cells, thus providing a possible link between the molecular pathophysiology of DM1 and visible histopathological alterations [56]. Additionally, Renard et al. proposed increased burden due to microvascular changes and lack of drainage of interstitial fluid and degraded protein products as a mechanism for anterior temporal white matter lesions in DM1 [58]. A key aspect of the present study is that the 1H-MRS VOI was placed for most of the patients within normal-appearing white matter (NAWM). The metabolic alterations detected by MRS are therefore present also in the NAWM, suggesting that they are either not related to the structural alteration of the tissue or appear before the morphological tissue damage. In fact, the proportion of normal-appearing and altered white matter within the investigated tissue by 1H-MRS were evaluated and did not correlate with the metabolites content suggesting that the white matter metabolic alterations in DM1 patients were not associated with morphological alterations of the same volume of interest. Widespread involvement of the NAWM in DM1 patients had already been shown by diffusion MR studies, that found altered diffusivity parameters also within brain regions not affected by the hyperintensities [5]. Our results support this evidence, also from a metabolic perspective. Overall, our results are in line with the reduction of NAA levels that was described in previous MRS studies in gray matter and white matter of DM1 patients. In particular, lower NAA level was shown in patients with congenital and juvenile myotonic dystrophy, supporting the hypothesis of a developmental disorder of neurons in the brain [9,13]. Lower NAA levels have been also shown at an early stage of the disease [11], and in both DM1 and DM2 patients [12]. More heterogeneous results are instead reported for the other metabolite content. Depletion of Cr and Cho levels, particularly in the frontal white matter, were found in DM1 patients and these changes in brain metabolites can differentiate DM1 and DM2, despite the similar structural MRI alterations, suggesting differences in their underlying pathophysiological mechanism [12]. Significant abnormalities in brain metabolites, especially higher mI and Cr, were also previously found, with a positive correlation with CTG repeat size [10]. Finally, evidence of brain (pathological accumulation of lactate) and skeletal muscle impairment of oxidative metabolism in DM1 patients has been also described through the MRS technique [14]. Previous proton MRS studies mainly focused on localization within the gray matter, such as the parietal cortex [9], midoccipital and temporo-parietal gray matter [10,12], insular cortex [11], anterior cingulate gyrus, and basal ganglia [13]. Only two studies also placed VOIs within the WM, focusing on the frontal WM [12,13]. The involvement of parieto-occipital regions in DM1 patients was previously demonstrated with morphological MRI [5,43] but no correlations with the neuropsychological data were reported. With the present results, we propose the parieto-occipital white matter as a crucial region of interest also for 1H-MRS, without necessarily including any WM lesions but focusing on the NAWM. Moreover, NAA alterations within POWM correlated with neuropsychological tests evaluating visuospatial and visuoconstructional functions, all features considered as recurrent in DM1 neuropsychological profile [16,17]. Specifically, positive correlations between NAA levels and neuropsychological functions showed that a decrease in NAA metabolites in parieto-occipital WM is bound to a worsening of the visuospatial and visuoconstructional performance. Moreover, a large part of our DM1 cohort presented a misperception by excess or by default (anosognosia) of the disease’s impact on their everyday life’s cognitive skills. Differently from the results presented by Baldanzi et al. [17], in our study, only a subgroup (8,82%) of our cohort presented a proper anosognosia. Those differences might be due to the different scales for anosognosia used in both studies, each one measuring different aspects of awareness. On the one hand, lack of disease awareness appears to be typically related to global brain atrophy in neurodegenerative disorders [59], and fMRI resting state studies suggested that self-awareness is supported by two closely interconnected neural networks, the “default mode network” (DMN) and the “attention system”, encompassing both frontal lobes and parietal structures [60,61]. On the other hand, visuospatial functioning is heavily reliant on the integrity of the parietal lobe and its connections with the occipital regions [62,63,64]. Neuroimaging studies in DM1 patients showed consistently a substantial impairment in visuospatial and visuoconstructional abilities that were significantly correlated with WM abnormalities and cortical volume loss as demonstrated by advanced morphological and diffusivity techniques such as voxel-based morphometry VBM and TBSS [5,18,26,65,66,67,68]. Here we showed that these abilities are also related to metabolic alterations that can be present within the NAWM. We focused on the WM involvement, from the metabolic and morphologic point of view, but, although DM1 patients have a widespread distribution of atrophy, a contribution from specific grey matter regions cannot be excluded, consistently with previous neuroimaging studies [5,18,67,68]. In conclusion, white matter metabolic changes highlighted by MR spectroscopy technique in adult DM1 patients while not being expression of tissue structural damage can be related to specific neuropsychological deficits. | true | true | true |
|
PMC9600469 | Vasileios Siokas,Polyxeni Stamati,Georgia Pateraki,Ioannis Liampas,Athina-Maria Aloizou,Daniil Tsirelis,Anastasia Nousia,Markos Sgantzos,Grigorios Nasios,Dimitrios P. Bogdanos,Efthimios Dardiotis | Analysis of SOD2 rs4880 Genetic Variant in Patients with Alzheimer’s Disease | 21-09-2022 | AD,Alzheimer’s disease,SOD2,oxidative stress,variant,rs4880,genetics | A few gene loci that contribute to Alzheimer’s Disease (AD) onset have been identified. Few studies have been published about the relationship between SOD2 rs4880 single nucleotide variant and AD, revealing inconsistent results. Therefore, the aim of the current study is to further examine the role of the SOD2 rs4880 in AD. We performed a case-control study with a total of 641 subjects (320 patients with probable AD, and 321 healthy controls). The statistical analysis was performed assuming five genetic models. The threshold for statistical significance was set at 0.05. The results revealed no association between SOD2 rs4880 and AD in any of the assumed genetic models that were examined [log-additive OR = 0.95 (0.76–1.19), over-dominant OR = 1.15 (0.85–1.57), recessive OR = 0.85 (0.59–1.22), dominant OR = 1.03 (0.72–1.47), and co-dominant OR1 = 1.10 (0.75–1.60) and OR2 = 0.90 (0.58–1.40)]. Adjustment for sex and subgroup analyses based on sex did not reveal any statistically significant results either. Based on our findings, SOD2 rs4880 does not appear to play a determining role in the risk of developing AD. Larger studies are warranted to elucidate the connection between rs4880 and AD. | Analysis of SOD2 rs4880 Genetic Variant in Patients with Alzheimer’s Disease
A few gene loci that contribute to Alzheimer’s Disease (AD) onset have been identified. Few studies have been published about the relationship between SOD2 rs4880 single nucleotide variant and AD, revealing inconsistent results. Therefore, the aim of the current study is to further examine the role of the SOD2 rs4880 in AD. We performed a case-control study with a total of 641 subjects (320 patients with probable AD, and 321 healthy controls). The statistical analysis was performed assuming five genetic models. The threshold for statistical significance was set at 0.05. The results revealed no association between SOD2 rs4880 and AD in any of the assumed genetic models that were examined [log-additive OR = 0.95 (0.76–1.19), over-dominant OR = 1.15 (0.85–1.57), recessive OR = 0.85 (0.59–1.22), dominant OR = 1.03 (0.72–1.47), and co-dominant OR1 = 1.10 (0.75–1.60) and OR2 = 0.90 (0.58–1.40)]. Adjustment for sex and subgroup analyses based on sex did not reveal any statistically significant results either. Based on our findings, SOD2 rs4880 does not appear to play a determining role in the risk of developing AD. Larger studies are warranted to elucidate the connection between rs4880 and AD.
Alzheimer’s disease (AD) is the most common form of dementia, as it constitutes 60–70% of dementia cases [1,2]. Human aging is a strong predisposing factor for AD with several cognitive, anatomical, and neurophysiological changes occurring with increasing age [3]. In fact, aging is strongly associated with AD, in a way that numerous people will manifest characteristics of the disease, and so the increasing age will inevitably increase the prevalence of AD [4]. AD is considered a neurodegenerative disease with chronic progression, starting with impairments in the ability to encode and save new memories [5]. Regarding the epidemiology of AD higher rates are reported in southern European countries compared to northern European countries, while women show a higher prevalence compared to men [6]. Inevitably, the data about the prevalence of people in preclinical stages of AD are scarcer, although 15–20% of people older than 64 years have Mild Cognitive Impairment (MCI), 14.9% of which are going to develop dementia in a couple of years [7]. A few genetic loci have already been identified that contribute to AD onset. However, it is necessary to identify new genetic risk factors via genome-wide association and candidate gene association studies, which have already revealed the complex nature of the genetics involved in AD [8]. Early-onset AD is categorized in most cases as early-onset familial Alzheimer’s disease (EOFAD), developed at a young age (<65 years), and represents 1–5% of total AD cases [9,10]. Other than the familial form, the sporadic form is responsible for 95% of AD cases [11,12], with the ε4 allele of the APOE gene exhibiting the strongest genetic association with sporadic AD [11,13]. Epigenetic mechanisms and environmental factors may also play a potential role in AD [11,14,15]. AD’s pathophysiological mechanisms have not been fully elucidated, though they mostly encompass amyloid-beta peptide accumulation in brain tissues and changes in the phosphorylation of Tau proteins in neurons which cause cytoskeletal impairments [16]. Moreover, many reviews refer to oxidative stress as an important contributor to the development of different neurodegenerative diseases, AD included [17]. The human brain is extremely sensitive to oxidative stress due to its elevated rate of oxygen metabolic activity, high levels of polyunsaturated fatty acids, a very oxidizable substrate, and low amounts of antioxidant systems compared to other organs [18,19]. Oxidative stress aggravates the production and accumulation of toxic amyloid plaques, as well as the phosphorylation of tau protein, the hallmarks of the pathogenesis of AD [20]. In addition, mitochondrial stress, and stress of the endoplasmatic reticulum (ER) have also been related to the pathophysiology of AD. Superoxide dismutase 2 (SOD2) is a manganese-containing enzyme belonging to the major antioxidant defense system of the organism. It is synthesized in the cytoplasm and is directly transported to the mitochondria where it protects the cell by catalyzing O2- generated by the respiratory chain, into H2O2 and molecular oxygen [21]. It possesses an essential role in maintaining mitochondrial proper function ensuring cellular survival [21,22]. The SOD2 gene is located in chromosome 6 in the region q25 [23,24]. In the available bibliography, studies have so far explored the relationship between the SOD2 rs4880 genetic variant and AD, and have yielded inconsistent results [25,26,27]. Therefore, the aim of the current study is to further examine the role of the SOD2 rs4880 in AD.
We conducted a case-control study in order to examine the effect of the SOD2 rs4880 genetic variant on the risk of AD. Written informed consent was granted from all individuals (or their close relatives when necessary) that were included in the present study. The Ethics Committee of the University General Hospital of Larissa approved the research protocol (132/17-06-2015).
The current case-control study includes the same sample from previously published articles [28,29,30,31]. In brief, we prospectively recruited patients with AD from the outpatient and inpatient clinics of the Neurology Department of the General University Hospital of Larissa, a tertiary referral institution located in Central Greece. The National Institute of Neurological and Communicative Disorders and Stroke/Alzheimer Disease and Related Disorders Association (NINCDS/ADRDA) criteria [32] were applied by a Senior Neurologist, in order for the diagnosis of probable AD to be set. The control group consisted of healthy volunteers with normal mini-mental state examination (MMSE) scores that did not fulfill the criteria for MCI and without any reported medical history record.
Leucocytes derived from peripheral blood samples from each participant were used, for nuclear DNA extraction. For this procedure, we applied the salting out method. After the isolation of the genetic material, we proceeded to genotype all the samples for the SOD2 rs4880 genetic variant. For the genotyping implementation, we applied the TaqMan allele-specific discrimination assay (Thermo Fisher Scientific, Waltham, MA, USA) on an ABI PRISM 7900 Sequence Detection System, while the SDS software (SDS 2.4) (Applied Biosystems, Foster City, CA, USA) was used to analyze the results. The entire method (PCR steps, enzyme activation, denaturation, and annealing/extension) has been previously extensively described [33]. Τo minimize any potential bias, the genotyping was carried out exclusively by personnel who were blinded to the clinical status of the samples.
Aiming to enhance the quality of the genotyping results and provide robustness, we re-genotyped a randomly selected 10% of the samples, leading to a 100% concordance with the initial genotyping results. Additionally, we set the threshold of the genotypic call rate (percentage of successfully genotyped samples) at 95%. Finally, we examined the genotyping results for any deviation from the Hardy–Weinberg Equilibrium (HWE), via Chi-squared test calculation [34].
The outcome of the current study was to explore any potential association between the SOD2 rs4880 and AD. The demographic characteristics are described as means ± standard deviation (SD) for continuous variables, and as n and/or percentages for categorical variables. The statistical power of the sample size was calculated using the CaTS Power Calculator for Genetic Studies (Center for Statistical Genetics, University of Michigan, Ann Arbor, MI, USA). The total number (n) and the percentages for allelic and genotypic distribution, in AD cases, in healthy controls, and the whole sample have been calculated. The effect size of the association between the SOD2 rs4880 and the AD was expressed in the terms of the odds ratios (ORs) and their precision [95% confidence intervals (CIs)]. The implementation of this analysis was made with the application of the SNPStats software (https://www.snpstats.net/, accessed on 10 April 2022) [34], assuming the following five genetic models (co-dominant, dominant, recessive, over-dominant, and log-additive). Subgroup analysis based on sex was also performed. The ‘T’ was considered as the reference allele and the ‘C’ as the alternative one for all of the analyses, with the exception of the female subgroup analysis where the “C” was considered as the reference. The p-value < 0.05 was set as the statistically significant threshold.
We genotyped an initial cohort of 654 subjects (327 patients with AD and an equal number of healthy controls). The overall genotype call rate was 98.01% (641/654). After excluding the subjects that failed to be genotyped, our cohort consisted of 320 patients with AD (33.8% male, mean age of blood collection ± SD = 78.88 ± 8.63 years) and 321 healthy controls (55.8% male, mean age ± SD = 69.72 ± 3.03). The basic demographic and clinical characteristics of the AD and healthy control groups are presented in Table 1. Moreover, we did not detect any deviation from the HWE in either AD patients or the healthy controls (p = 0.37 and p = 0.82, respectively). Our sample had a power of 80.3 to detect a significant association (p < 0.05) between the SOD2 rs4880 variant and AD, with the frequency of the C allele equal to 50%, a prevalence of AD equal to 37/100,000, and a relative risk of 1.37 for the multiplicative mode of inheritance. The allelic distribution for the minor allele (G) was 49% in the group of patients with AD, and 50% in the healthy controls group. Moreover, the genotypic distribution was 79 (25%), 169 (53%), and 72 (22%) for the T/T, T/C, and C/C in the group of patients with AD. The values in the healthy control group were 81 (25%), 158 (49%), and 82 (26%), respectively. The SOD2 rs4888 variant total numbers of the alleles and genotypes for all subjects, the healthy controls, and for the patients with AD are presented in Table 2. No association (p > 0.05) was found between the SOD2 rs4880 and AD in any of the examined genetic models of inheritance [log-additive OR = 0.95 (0.76–1.19), over-dominant OR = 1.15 (0.85–1.57), recessive OR = 0.85 (0.59–1.22), dominant OR = 1.03 (0.72–1.47), and co-dominant OR1 = 1.10 (0.75–1.60) and OR2 = 0.90 (0.58–1.40)]. Adjustment for sex could not reveal any statistically significant results (p > 0.05) [log-additive OR = 0.91 (0.73–1.15), over-dominant OR = 1.15 (0.83–1.58), recessive OR = 0.81 (0.56–1.16), dominant OR = 0.98 (0.68–1.42), and co-dominant OR1 = 1.05 (0.71–1.55) and OR2 = 0.84 (0.53–1.22)]. The respective results are depicted in Table 3. Subgroups analysis based on sex revealed no association (p > 0.05) between rs4880 and AD. The SOD2 rs4888 variant total numbers and frequencies of the alleles and genotypes based on sex are depicted in Table S1 (see Supplementary material) for male participants and in Table S2 for female participants. Results regarding the association analysis between rs4880 and AD are presented in Table S3 for the male participants and in Table S4 for the female participants.
In the present case-control study, we recruited a large number of patients with AD as well as healthy controls and genotyped them for the SOD2 rs4880 variant. We did not detect any association between this variant and the risk of developing AD. To the best of our knowledge, this is the first study to investigate this variant and its possible correlation with AD in a sample of Greek origin. The SOD2 gene is located in chromosome 6 in the region q25 [23,24]. Almost 190 genetic variations variants have been described in the SOD2 gene until now, and have been involved in breast cancer, diabetes mellitus, dyslipidemia, and other diseases [24]. The SOD2 gene encodes the SOD2, a manganese-containing enzyme that belongs to the major antioxidant defense system. Impaired SOD2 enzymatic activity leads to increased oxidative stress [35]. The rs4880 (C47T) is a missense variant in exon 2 located at Chromosome 6:159692840 (http://www.ensembl.org/Homo_sapiens/Variation/Explore?r=6:159692340159693340;v=rs4880;vdb=variation;vf=167107369, accessed on 10 April 2022). When the T allele is present, the amino acid valine (Val) is encoded at codon 16, while the amino acid alanine (Al) is encoded instead in the case of the C allele [36,37]. The T allele of the SNP rs4880 results in structural alterations in the mitochondrial targeting domain of SOD2, leading to its less efficient post-transcriptional transport into the mitochondrion and decreased potential in neutralizing superoxide anions [37]. Studies have so far explored the relationship between SOD2 rs4880 and AD have yielded inconsistent results [25,26,27]. The study of Wiener et al. (2007) revealed significant indication of an association between rs4880 and AD, using family-based association tests. [25]. The case-control study of Spisak et al. (2014) did not prove any association between rs4880 and AD [26]. Finally, the study of Gamarra et al. (2015) showed that the SOD2 rs4880 in combination with APOEε4 allele carriage increases the risk for MCI, while it also increases the risk for AD compared to MCI [27]. Oxidative stress represents an imbalance between oxidants and antioxidants in favor of the oxidants, leading to severe damage to proteins, lipids, DNA, and RNA of the cells at several levels [38]. The main representatives of oxidative stress are the reactive oxygen species (ROS) and reactive nitrogen species (RNS), products of normal cellular metabolism, which can play both beneficial and catastrophic roles to the cell and the organism [39,40]. In low amounts, they act as a part of the immune system against infectious agents, contribute to the cellular signaling system, and activate several important metabolic pathways of the cellular metabolism [39,40]. On the other hand, when they are found in excessively high amounts, as in oxidative stress, they cause severe biologic damage inducing cellular death [41]. Mitochondria, as a major source of ROS production, are significantly vulnerable to oxidative stress and their malfunction contributes to the course of aging and neurodegenerative processes [20,42,43]. Many studies have shown a number of changes in the mitochondria in the brains of patients with AD, such as smaller size, lower number, impaired mitochondrial transport to regions of high energy demands, and altered mitochondrial fission–fusion protein dynamics, all leading to loss of synaptic plasticity and normal neuronal function, contributing to neurodegeneration [44,45,46]. AD represents a major example of the protein folding disease pattern. The ER is the center of folding and transport of newly formed proteins in the cell membrane, while it also keeps a key role in the maintenance of calcium homeostasis [47]. Any disturbance in the normal protein folding process may lead to the accumulation of misfolded or unfolded proteins in the ER. In response to that, an important pathway called the unfolded protein response (UPR) is activated in order to protect the cell [48]. Many studies have confirmed that UPR is involved in the early stages of AD, and ER stress markers have been described in tissues derived from patients with AD. Additionally, a self-implying relationship has been shown between hyperphosphorylation of tau protein and UPR activation, leading to a vicious cycle of neuronal degeneration [49,50]. Similarly, toxic aβ oligomers, through a drastic disturbance of ER calcium homeostasis, produce even more altered protein folding and oxidative stress, leading to cell death [51]. For this important role, studies have shown that SOD2 deficiency causes mitochondrial damage in cells with high levels of oxidative metabolism, like neurons, hemopoietic, and hepatic cells [52]. Impaired SOD2 activity has been related to several diseases such as cancer, particularly ovarian and breast cancer, diabetes mellitus, and neurodegeneration, including AD [52,53]. It seems that SOD2 is an important factor in the pathogenesis of the disease since it has a key role in repairing oxidative damage. Additionally, SOD2 may regulate neuroinflammation by controlling the activation of microglia. Interestingly, studies indicate that at the beginning of the inflammatory response activated microglia regulate the activation of SOD2 and at the end of inflammation increased SOD2 inactivates microglia, protecting the cell from the oxidative stress of chronic inflammatory processes [54].S Moving on, we must acknowledge the limitations of the present study. Firstly, some of the patients’ data have been based on patients’/caregivers’ self-reporting, as most of them were hospitalized with advanced AD stage. Therefore, it was not possible to include the precise age at AD onset in the analyses. Consequently, we performed adjusted analyses only for sex without including additional potential predisposing or precipitating AD risk factors (genetic and not-genetic, especially the APO ε4 carriage status) in regression statistical models. Therefore, the possibility that our results are significantly affected by the latent effect of uncontrolled co-founders cannot be totally excluded. This could partially explain the lack of an association in our study, as the SOD2 rs4880 variant may be an additional risk factor for the development of an amnestic syndrome. This was reciprocated in the study of Gamarra et al. (2015), where SOD2 rs4880 T allele carriage was shown to increase the risk of amnestic MCI patients carrying APOEε4 [27]. Following the previous limitation, we included subjects without screening for major AD-linked genes [8]. Finally, additional analyses correlating SOD2 rs4880 genotyping data with other phenotypes (e.g., age at AD onset, disease duration, disease progression, and MMSE score) would have provided more robustness to our conclusions.
In conclusion, we provide these first data from a Greek population regarding the SOD2 rs4880 variant and AD. It is crucial that additional studies be performed in order to elucidate the role of SOD2 rs4880 in AD. More precisely, large multicenter studies, possibly also measuring the enzymatic activity of SOD2 and also adjusting for co-founders that confer susceptibility to AD in order for the attributable risk of this variant to AD to be fully elucidated are needed. | true | true | true |
|
PMC9600495 | Marcella R. Cardoso,Alex Ap. Rosini Silva,Maria Cecília R. Talarico,Pedro H. Godoy Sanches,Maurício L. Sforça,Silvana A. Rocco,Luciana M. Rezende,Melissa Quintero,Tassia B. B. C. Costa,Laís R. Viana,Rafael R. Canevarolo,Amanda C. Ferracini,Susana Ramalho,Junier Marrero Gutierrez,Fernando Guimarães,Ljubica Tasic,Alessandra Tata,Luís O. Sarian,Leo L. Cheng,Andreia M. Porcari,Sophie F. M. Derchain | Metabolomics by NMR Combined with Machine Learning to Predict Neoadjuvant Chemotherapy Response for Breast Cancer | 15-10-2022 | 1H-NMR,breast neoplasms,magnetic resonance spectroscopy,metabolism,untargeted metabolomics,drug resistance | Simple Summary Neoadjuvant chemotherapy (NACT) is offered to breast cancer (BC) patients to downstage the disease. However, some patients may not respond to NACT, being resistant. We used the serum metabolic profile by Nuclear Magnetic Resonance (NMR) combined with disease characteristics to differentiate between sensitive and resistant BC patients. We obtained accuracy above 80% for the response prediction and showcased how NMR can substantially enhance the prediction of response to NACT. Abstract Neoadjuvant chemotherapy (NACT) is offered to patients with operable or inoperable breast cancer (BC) to downstage the disease. Clinical responses to NACT may vary depending on a few known clinical and biological features, but the diversity of responses to NACT is not fully understood. In this study, 80 women had their metabolite profiles of pre-treatment sera analyzed for potential NACT response biomarker candidates in combination with immunohistochemical parameters using Nuclear Magnetic Resonance (NMR). Sixty-four percent of the patients were resistant to chemotherapy. NMR, hormonal receptors (HR), human epidermal growth factor receptor 2 (HER2), and the nuclear protein Ki67 were combined through machine learning (ML) to predict the response to NACT. Metabolites such as leucine, formate, valine, and proline, along with hormone receptor status, were discriminants of response to NACT. The glyoxylate and dicarboxylate metabolism was found to be involved in the resistance to NACT. We obtained an accuracy in excess of 80% for the prediction of response to NACT combining metabolomic and tumor profile data. Our results suggest that NMR data can substantially enhance the prediction of response to NACT when used in combination with already known response prediction factors. | Metabolomics by NMR Combined with Machine Learning to Predict Neoadjuvant Chemotherapy Response for Breast Cancer
Neoadjuvant chemotherapy (NACT) is offered to breast cancer (BC) patients to downstage the disease. However, some patients may not respond to NACT, being resistant. We used the serum metabolic profile by Nuclear Magnetic Resonance (NMR) combined with disease characteristics to differentiate between sensitive and resistant BC patients. We obtained accuracy above 80% for the response prediction and showcased how NMR can substantially enhance the prediction of response to NACT.
Neoadjuvant chemotherapy (NACT) is offered to patients with operable or inoperable breast cancer (BC) to downstage the disease. Clinical responses to NACT may vary depending on a few known clinical and biological features, but the diversity of responses to NACT is not fully understood. In this study, 80 women had their metabolite profiles of pre-treatment sera analyzed for potential NACT response biomarker candidates in combination with immunohistochemical parameters using Nuclear Magnetic Resonance (NMR). Sixty-four percent of the patients were resistant to chemotherapy. NMR, hormonal receptors (HR), human epidermal growth factor receptor 2 (HER2), and the nuclear protein Ki67 were combined through machine learning (ML) to predict the response to NACT. Metabolites such as leucine, formate, valine, and proline, along with hormone receptor status, were discriminants of response to NACT. The glyoxylate and dicarboxylate metabolism was found to be involved in the resistance to NACT. We obtained an accuracy in excess of 80% for the prediction of response to NACT combining metabolomic and tumor profile data. Our results suggest that NMR data can substantially enhance the prediction of response to NACT when used in combination with already known response prediction factors.
Breast cancer (BC) is the most common cancer worldwide [1] and the leading cause of cancer-related mortality among women in Brazil [2]. Although BC presents histological similarity, clinically it can be very heterogeneous, with several phenotypic and genotypic subtypes [3], which results in a challenge in treatment effectiveness. Molecular and immunohistochemical markers generally are used to classify BC subtypes. The classification is based on the expression of hormonal receptors (HR) such as estrogen and progesterone receptors (ER; PR), human epidermal growth factor receptor 2 (HER2), and the cell proliferation marker (nuclear protein Ki67). These markers are also related to therapeutic decisions and prognostics [4,5]. One of the treatments used prior to definitive surgical therapy is neoadjuvant chemotherapy (NACT), which has gained attention due to its ability to reduce tumor size and cancer burden, avoiding mutilating surgical procedures [6]. It also has the potential for increasing resectability and controlling the micrometastatic disease. NACT provides a viable alternative when there is poor radiotherapy access or there are unavoidable delays in delivering radiotherapeutic treatment. To evaluate whether the patient had a favorable response after completion of NACT, pathological complete response (pCR) is defined as the absence of invasive tumors in the breast and axilla. This response is associated with improved long-term survival rates or far lower risk of subsequent recurrence [7]. BC patients who achieved a pCR after the NACT showed higher rates of disease-free survival (DFS) and overall survival (OS) than women with residual disease in their surgical specimens obtained after NACT [8]. Unfortunately, about 30% of patients show a pCR to NACT. Patients with HR-positive tumors have a lower probability of pCR after NACT than patients with HER2-positive and triple-negative tumors [9]. The well-known HR, HER2, as well as the Ki67 markers triad, explain a substantial portion of the NACT response variability. Nevertheless, a better understanding of the mechanisms that define the response to NACT is still needed to determine which patients will gain clinical benefits from NACT. Many patients undergoing NACT have the adverse effects of therapy without enjoying pCR, and oncologists are largely incapable of predicting unsuccessful outcomes. To further improve the knowledge of the molecular events leading to pCR after NACT, we designed and carried out the present metabolomic study to construct a biomolecular and metabolic portrait of the biochemical processes in fluids and tissues of BC patients who had undergone NACT [10,11,12]. Metabolites are final byproducts acquired from the interaction between intracellular pathways and their microenvironment [13]. Metabolomics, when faced with a stimulus such as a disease, measure a comprehensive set of metabolites, thus representing the bioactivity of a system [14]. Thus, in cancer, metabolomics detects oncological developments by evaluating measurable metabolic profiles and selecting metabolic pathways through global variations in metabolites [15,16]. Specifically for breast cancer, several articles have proposed that assessing a metabolite profile might allow for an understanding of the biochemical processes that occurred or were occurring at the time of diagnosis [17,18,19]. In addition, in the chemoresistance field, metabolomics allows for the individual characterization of the patient, enabling the personalization of treatment with strategies focused on maximizing the action of drugs. Thus, metabolomics can be developed as a sensitive prognostic tool for associated therapy for diseases such as breast cancer [20,21]. However, a few studies have evaluated metabolites associated with better response to NACT using metabolomics [22,23,24]. The use of bioinformatics allows the use of approaches for predictive modeling, using medical data, for cancer evaluation [25]. For example, machine learning is able to generate models for prediction by testing vastly through model and parameter space, as opposed to traditional statistics approaches, which have a limited set of hypotheses [26,27,28,29,30]. In our study, we analyzed the serum metabolites of patients with different molecular subtypes of BC who had undergone NACT using Nuclear Magnetic Resonance (NMR). We also created machine learning classifiers by correlating the observed metabolites with the expression of BC markers and creating models to predict the response to NACT. These models were able to predict the response to NACT using pre-treatment serum samples.
This is a prospective cohort study on 80 women aged between 29 and 77 years with invasive breast carcinoma who underwent NACT followed by surgery. Participants were diagnosed and treated at the Women’s Hospital (Hospital da Mulher Prof. Dr. José Aristodemo Pinotti, Centro de Atenção Integral à Saúde da Mulher–CAISM) of the University of Campinas (UNICAMP) in Brazil between January 2017 and January 2019. Demographic and clinical data from patients’ records are summarized in Table 1. The biological samples were stored in the CAISM’s biobank (CONEP 56, Campinas, SP, Brazil). All study subjects signed informed consent forms before removing their biological samples and being included in the institutional biobank. The Research Ethics Committee approved the study of UNICAMP (CAAE, 69699717.0.0000.5404).
Tissue samples were collected with an ultrasound-guided percutaneous needle (core) biopsy either at biopsy (pre-treatment specimens) or during surgical resection after NACT (post-treatment specimens). The samples were formalin-fixed and paraffin-embedded. Hematoxylin-eosin-stained sections were reviewed to confirm the histologic diagnosis. NACT was prescribed according to the standard protocols, including doxorubicin and cyclophosphamide, followed by paclitaxel (with carboplatin in triple-negative cases). HER2-positive patients received trastuzumab. See Supplementary Tables S1 and S2 for the detailed information of the therapeutic regimen. After NACT, all women underwent surgical treatment (mastectomy or quadrantectomy with sentinel lymph node biopsy or axillary lymph node dissection). All histological diagnoses were determined according to the World Health Organization (WHO) criteria and following the grade as per the Nottingham classification (Supplementary Methods) [31].
For BC subtype classification, a conventional manual immunohistochemical technique was used [31,32,33,34,35]. Anti-estrogen receptor (ER, clone 1D5, diluted at 1:1000 v/v), anti-progesterone receptor (PR, clone PR636, diluted at 1:800 v/v), anti-HER2 (Clone PN2A, diluted at 1:1100 v/v), and anti-Ki67 (clone MIB1, diluted at 1:500 v/v) were used as primary antibodies (all by Dako, Agilent, Santa Clara, CA, USA). Evaluation of ER and PR was performed accordingly, on the diagnostic (pre-treatment) specimens [32]. Cases were considered positive when ≥1% of tumor cells were positive. For Ki67, we scored the percentage of positive tumor cells in a minimum of 500 cells in randomly selected representative fields [33]. When the mean percentage of stained cells was higher than the median (40%), cases were assigned as Ki67 high. Human epidermal growth factor receptor 2 (HER2) staining was scored as 0+/1+ (negative), 2+ (equivocal), or 3+ (positive). Equivocal (2+) cases were further confirmed by in situ hybridization, according to the recommendations of the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) [34]. Due to tumor heterogeneity, hormone-receptor-negative cases and/or HER2-negative had the immunohistochemistry repeated in the surgical specimen of women with a residual disease for subtype confirmation [35].
The response to NACT was evaluated on surgical specimens removed after NACT according to the Residual Cancer Burden (RCB) guidelines [36,37]. For statistical analyses, all cases were clustered into two major groups based on the response to NACT: (a) cases with a pCR and RCB-I (minimum residual disease) as sensitive, and (b) cases with residual breast carcinoma RCB-II (moderate residual disease) and RCB-III (extensive residual disease) as resistant [38].
Aliquots of peripheral blood were collected in dry tubes (Vacuette® tube 2.5 mL CAT Serum Separator Clot Activator 13 × 75 red cap-white ring, non-ridged, North Carolina, USA) and obtained from women before they initiated the NACT. Immediately after blood collection, the tubes were mixed by inverting several times and incubated for ≥30 min at 4 °C. The tubes were centrifuged for 10 min at 1000× g. The samples were stored at −80 °C until analysis. Serum samples were thawed at room temperature before the spectroscopy analysis. Then, 400 μL of serum was slowly mixed with 200 μL of D2O (99.9% deuterium oxide with 0.03% of TSP) and transferred to 5 mm NMR tubes. The proton NMR spectra (1H-NMR) were acquired at 298 K using a Varian Inova® of 599.887 MHz NMR spectrometer (Agilent Technologies® Inc., Santa Clara, CA, USA) equipped with a triple resonance cryoprobe. Regular one-dimensional 1H-NMR spectra were obtained using CPMG (Carr–Purcell–Meiboom–Gill) pulse sequence with ns = 128 with 8 K data points at a spectral sweep width of 16 ppm; total relaxation delay of 4 s; water presaturation applied during 2 s delay; T2 filtering was obtained with an echo time of 300 μs repeated 166 times, resulting in a total duration of effective echo time of 50 ms. The NMR acquisitions were performed at the Brazilian Biosciences National Laboratory (Brazilian Center for Research in Energy and Materials, CNPEM, Campinas, SP, Brazil).
The obtained spectra were processed using the MestreNova software (MestrelabResearch S.L.). The chemical shifts of the spectra were referenced using the doublet corresponding to the lactate’s methyl group (–CH3) at 1.324 ppm (3H, d, 3JHH = 7.0 Hz). The extent of information was reduced by using a binning of 0.005 ppm. The samples were normalized by the sum. The processed spectra were converted into data matrices to prepare them for identification and analysis. In addition, regions of the spectrum corresponding to substances with a potential for interference in the analyses, such as water (4.40–5.23 ppm), were removed. The NMR analyses were performed at the Biological Chemistry Laboratory from UNICAMP. Chenomx NMR® Suite 8.1 software (Chenomx® Inc., Edmonton, AB, Canada) was used to quantify relative concentrations.
All statistical calculations were performed using the R Foundation for Statistical Computing (v3.6.2), Vienna, Austria. The epitools package was used for the odds ratio (OR) calculation of clinical characteristics of the women according to their response to NACT. Confidence levels of 95% and p-values <0.05 were assumed as statistically significant [39]. In addition, we used standard methodology to calculate the sample size for the study. Assuming a 0.05 significance level, a Cohen coefficient of 0.8 (high), the power for our analysis sits at a comfortable 0.801, considering the proportion of NACT sensitive/resistant patients in the cohort of 80 women with complete clinical data available for analyses. The caret package was used for recursive feature elimination (RFE) and logistic regression (LR) modeling [40]. RFE was applied for continuous elimination of features (i.e., metabolites or clinical markers) with a low contribution to the model [41,42,43]. Nine classification prediction models for the response to NACT were generated based on the alternative combinations of HR, Ki67, HER2 statuses, and the metabolite panel (Supplementary Table S3). We performed separate analyses for the entire (Supplementary Table S3) and triple-negative/HER2+ tumors (n = 45, Supplementary Table S4). Models were built based on a training set comprising 75% of the data. To evaluate the performance of the models, we performed Leave-One-Out Cross-Validation (LOOCV). The area under the curve (AUC) of the receiver operator characteristics (ROC) curve of the selected features was also used to evaluate their prediction power. A validation set comprising 25% of data was also used for performance evaluation in terms of sensitivity, specificity, and accuracy [44]. For metabolomics analysis, normalization by sum and pareto scaling data were performed using the web platform MetaboAnalyst™ 5.0 [45]. Using this platform, we further interrogated the KEGG Library for quantitative Metabolite Pool enrichment, and the pathways with a p-value < 0.05 were considered significant.
Eighty patients included in this study and undergoing NACT presented invasive ductal carcinoma (100%), most of which were histological grade 3 (51%), ER-positive (71%), and PR-positive (65%). Patients with tumors of histological grade 3 had a higher probability of pCR/RCB-I (OR = 5.57 (1.439–21.470); p = 0.0161) than their counterparts whose tumors were grade 1/2. In contrast, women with HR-positive tumors were less likely to enjoy pCR/RCB-I (OR = 0.18 (0.04–0.670); p = 0.005) than the women with non-luminal tumors. Table 2 shows the distribution of women according to the disease characteristics and NACT response.
Untargeted NMR analysis identified and quantified the relative abundances of more than 27 compounds, including 14 amino acids (arginine, asparagine, glutamate, histidine, leucine, lysine, phenylalanine, proline, serine, threonine, tyrosine, and valine) and other metabolites (Table S5). Overall, no significant differences in the number of metabolites identified by 1H-NMR in the spectra of resistant and sensitive patients were observed. Representative spectra obtained from serum patients in the resistant and sensitive groups are shown in Figure 1. Univariate analysis shows that leucine and formate were found to be significantly altered when comparing resistant and sensitive women (p-value < 0.05) (Figure 2 and Table S5); however, their predictive power evaluated using a logistic regression model showed to be unsatisfactory (AUC = 0.67) (data not shown).
Logistic regression (LR) models have supported the assessment of biomarkers in cancer [46,47,48,49] and can also be coupled with RFE [50], as we present in Figure 3. When compared to the univariate approach, we observed an increase in the predictive power for the model using only the abundance of the metabolites (AUC = 0.83 – Model I, Figure 3A). Using the LR-RFE, we found formate, proline, valine, and leucine as potential markers of NACT response (Figure 3B). By combining the abundances of these metabolites with the information on HR status (Model II), HR and HER2 (Model VI), and Ki67, HER, and HR (Model VIII), we obtained AUC values higher than those obtained for the model that solely used metabolites. The contributions of the metabolites were found to be more significant for the models than the contribution of the molecular markers, as presented by the observed coefficient values (Figure 3B). For the total contribution of the HR, Ki67, HER2, and serum metabolites as predictors of the response to NACT in the different models obtained, see Table S6. The coefficient values display the patient’s chance of being resistant to NACT. A positive value means a positive correlation with NACT resistance, whereas a negative value means a positive correlation with NACT sensitivity. For example, formate, proline, valine, HR+, and HER2− display positive coefficient values, thus being directly related to the NACT resistance. It means that once any of these predictors show an increase, the patient’s chance of being resistant to NACT increases. Conversely, leucine, HR−, and HER2+ have negative coefficient values and are inversely related to NACT resistance or directly associated with NACT sensitivity. It means that an increase in leucine concentration, keeping the remaining predictors constant, increases the patient’s chance of being sensitive to the treatment, in accordance with what was previously pointed out by the univariate analysis of the metabolite set. Ki67, both positive and negative, was found to have a low influence on the models that contain it. When exploring the performances of the models based on their predictive power and agreement with the pathologic results, equivalent specificity rates were found for the eight models when considering the training set (Supplementary Table S3). For the validation sets, the specificity reached 75% only for the models II and V (Figure 3C), which were considered the ones with the highest efficiency in the classification. The sensitivity of the models, for both training and validation sets, was higher than 70% for all the models (Supplementary Table S3). The results suggest that the interconnection of clinical variables (HR/Ki67/HER2) and serum metabolites can improve the prediction of the response to NACT. When the analyses were performed in a restricted cohort of patients with triple-negative and HER2+ tumors, similar performance estimators were obtained (Supplementary Table S4). To gain insights into metabolome changes associated with resistance and sensitivity to NACT, the pathway enrichment analysis was carried out with the metabolites found as discriminants for models I–VIII. Glyoxylate and dicarboxylate metabolism was the most enrolled pathway for acquiring resistance (Supplementary Figure S1, Supplementary Materials).
The potential of the combination of HR, Ki67, and HER2 statuses and serum metabolites in predicting the response to NACT was evaluated in this study. The untargeted NMR analysis identified and quantified the relative abundances of 28 compounds, including 14 amino acids, carnitine, and other metabolites. Amino acids are an indispensable source of nutrients for all types of cells. Both essential and non-essential amino acids have an important role in providing build blocks for cell growth and proliferation. The critical role of these amino acids is even more important for tumor cells due to their rapid growth and proliferation, supporting protein synthesis [51]. It is also known that amino acids cannot cross the cell membrane without the assistance of specific transporters due to their hydrophilic component [52]. Therefore, since tumor cells require a high demand of amino acids to satisfy their rapid growth and proliferation, amino acid transporters’ expression is higher than normal cells [53]. Furthermore, the amino acid transporters are expressed in different levels in types of cancers and exhibit different properties in substrate selectivity [54]. When studying the serum abundance of this set of metabolites, we noticed the significant differences between the concentrations of leucine and formate. Leucine, which has been extensively studied for its role in breast cancer, showed a lower concentration in the serum of resistant women [55,56,57]. Saito et al., 2019, proposed that resistance is based on the increased expression of transporters that incorporate leucine to fuel the accelerated proliferation of cells resistant to therapy, which supports our observation of a lower concentration of this amino acid in resistant patients to NACT [58]. The higher concentration of leucine observed in the serum from sensitive in comparison to resistant patients could reflect less leucine uptake from sensitive tumors. One possible explanation could be a lower expression of L-type amino acids transporter 1 (LAT1) in sensitive tumors. Leucine is an essential amino acid [59] that activates the mammalian target of rapamycin (mTOR) which regulates cell growth and cell cycle progression [60], stimulating insulin signaling [61] or oxidized for energy purposes by tumors [62]. Therefore, within a lower leucine uptake, the mTOR pathway will be down-regulated, growth stimuli and energy source will be decreased, and consequently, less cell growth and proliferation could lead to the sensitive phenotype. Contrarily, Yang and collaborators (2018) performed a metabolomics study evaluating the plasma profile associated with response to neoadjuvant chemotherapy for colorectal cancer patients and found that leucine concentration was decreased in responsive patients [61]. Compared to Yang et al. (2018), this discrepancy in the present study could be explained by the heterogeneity among different cancers, such as cell type and histologic classification. A study by Saito and colleagues (2019) demonstrated that leucine imported by LAT1 was involved in tamoxifen resistance in ER-positive breast cancer [58]. In addition to leucine, another amino acid, arginine, had different serum profiles when comparing sensitive and resistant patients. Arginine is a semi-essential amino acid that could be derived primarily from diet and synthesis in the kidney [63]. Arginine is crucial in many biologic processes, including the immune system. Furthermore, arginine is essential to cellular growth and may become limited in cases of rapid proliferation such as malignancy [64]. Jayant and Anant (2017) evaluated the serum levels of arginine in breast cancer patients and found lower levels in cancer patients, independent of stage, compared to healthy controls [65]. In the present study, we found higher levels of arginine in resistant patients in comparison to sensitive ones. This result could reflect that the required arginine may differ between sensitive and resistant tumors. In addition to alterations in the serum amino acid profile, other metabolites such as formate have their profile changed among sensitive and resistant breast cancer patients. Formate is produced by many metabolic pathways such as serine catabolism, through the mitochondrial pathway. Sterol synthesis and tryptophan catabolism are also involved in this process. Formate may have different directions depending on cell type and environmental conditions. In proliferating cells, such as cancer, formate contributes to the one-carbon (1C) demand for the synthesis of nucleotides [66]. The high demand for nucleotide, RNA, and DNA is essential for rapid tumor growth [67]. For this reason, the circulating formate levels are reduced in some cancer patients, such as breast and lung cancer, relative to healthy controls [68]. Formate was increased in resistant women and can be considered a potential predictive biomarker, in agreement with several publications on cancer [22,68,69]. Jiang et al. (2018) performed a pharmacometabolomics study in metastatic breast cancer patients evaluating the response to gemcitabine–carboplatin chemotherapy. The authors found that significantly lower baseline levels of serum formate in patients with the resistant disease may reflect the higher demand from them for alternate/additional nutritional sources to fuel the accelerated proliferation of breast cancer cells biologically more aggressive or resistant to therapy [70]. Contrarily, our results revealed significantly higher serum formate in resistant patients. This controversial result could be explained by the fact that the oxidative nature of cancers and the chemotherapy treatment for our patients may differ from those included in the study performed by Jiang. Subsequently, we used the recursive feature elimination (RFE) method to assist in the selection of putative biomarkers. RFE continually removes the resources with low contribution scores based on the iterative method and then classifies each resource in each cycle to exclude the resources with low scores [43]. Studies have proposed that RFE allows the extraction of potential biomarker subsets among different cancer types [71,72,73]. Specifically, for breast cancer, RFE was applied to classify the complete pathological response and distinguish triple-negative breast cancer from other subtypes of breast cancer based on the selection of miRNA biomarkers [74,75]. Logistic regression (LR) models, following the RFE approach, are also widely used in classification problems applied to breast cancer [76,77,78,79]. In our study, RFE + LR was applied as a supervised learning machine by combining the abundance of metabolites with the clinical information on HR, Ki67, and HER2 status. The resulting classification models presented sensitivity and specificity values greater than 70% and 80%, respectively, for the training set. Remarkable performance was achieved for Model II, with 75% sensitivity and 83% specificity for the training set and 81% sensitivity and 75% specificity for the validation set. This indicates that combining the HR status with the metabolite panel (leucine/valine/formate/proline) can generate good classifiers of the response to NACT. A few studies reported similar results for predicting the response to NACT in BC patients based on metabolite panels [22,23,24]. Lin et al. (2019) assessed the metabolic biomarker signature of serum samples of BC patients by liquid chromatography–mass spectrometry (LC-MS). The authors applied partial least squares discriminant analysis (PLS-DA) to build the statistical model of classification, which identified nine informative metabolites and achieved performances with a specificity of 100% and sensitivity of 81.2% [23]. Some authors recently examined serum metabolite levels during chemotherapy treatment. Debik et al. (2019) observed unfavorable changes in lipid levels during NACT. By using NMR, they observed no metabolic difference in serum samples from survivors and non-survivors, although a PLS-DA model based on their analysis of tissue achieved 72% accuracy in predicting a 5-year survival rate. Among other metabolites, lactate, glycine, choline, and alanine were reported as the critical variables useful for the model set-up [22]. Vignoli et al. (2020) also studied metabolic profiles capable of codifying a complete response to NACT within the ER-positive group. After investigating the plasma by NMR, they built a classifier with low performance. Among other findings, branched-chain amino acids, such as valine and isoleucine, were reported as the key differentiators of their groups of study [24]. Unlike the previous studies mentioned above, applying the RFE + LR method allowed us to obtain models with successful results in predicting the response to NACT that links the selected biomarkers with the clinical information of patients. Glyoxylate and dicarboxylate metabolism was the pathway most impacted by the acquisition of resistance (Supplementary Figure S1). This pathway has been associated with breast cancer cell metastasis by gas chromatography–mass spectrometry (GC-MS) and direct infusion mass spectrometry [80]. In addition, other pathways represented by our panel of metabolites are associated with breast cancer, such as the metabolism of branched-chain amino acids (BCAAs, i.e., valine, leucine, and isoleucine), the Aminoacyl-tRNA, arginine and proline, Pantothenate, and CoA [50,81]. Together, this set of relationships highlights the metabolic alterations involved in the cellular reprogramming that provoked the BC resistance to chemotherapy. The current analysis is based on a strong dataset of comprehensive clinical and metabolomic data, derived from a well-maintained biobank, coupled with clinical facilities designed to collect world-class clinical data. We acknowledge, on the other hand, that a larger cohort of patients with non-luminal tumors would have enhanced the possibility of subset analyses aimed at refining the performance estimations on less common tumor types, such as triple-negative and HER2+ tumors. Although untreatable at this point, we are conducting further recruitment of patients with these less common tumor types.
In conclusion, NMR of serum offered rapid access to metabolic alterations in BC associated with resistance and sensitivity to NACT. The reported proof-of-principle study attested the power of these techniques in accelerating the definition of BC prognosis with excellent accuracy. Screening methods for predicting the response and effects of chemotherapies are much desired and would be of paramount importance for cancer patients. Despite BC being one of the most treatable cancers, a massive number of women do not benefit from NACT pre-treatment because of resistance to chemotherapy. Capitalizing on the close relationship between cancer formation and changes in serum, the analysis of the metabolites by NMR allowed insights into the metabolic alterations associated with marker status and resistance to NACT. When coupled by LR + RFE to the clinical marker status, a simple metabolomic panel can be applied successfully as an add-in prognosis and clinical forecasting for the response to NACT of BC patients. Although rigorous multisite validation of this untargeted approach with a larger patient pool is needed, this is the first milestone for developing an efficient strategy for the early discovery of NACT-resistant BC patients. | true | true | true |
|
PMC9600526 | Lei Dong,Yang Li,Liqun Liu,Xinyi Meng,Shengzhen Li,Da Han,Zhenyu Xiao,Qin Xia | Smurf1 Suppression Enhances Temozolomide Chemosensitivity in Glioblastoma by Facilitating PTEN Nuclear Translocation | 20-10-2022 | TMZ,drug resistance,glioblastoma,smurf1,PTEN | The tumor suppressor PTEN mainly inhibits the PI3K/Akt pathway in the cytoplasm and maintains DNA stability in the nucleus. The status of PTEN remains therapeutic effectiveness for chemoresistance of the DNA alkylating agent temozolomide (TMZ) in glioblastoma (GB). However, the underlying mechanisms of PTEN’s interconnected role in the cytoplasm and nucleus in TMZ resistance are still unclear. In this study, we report that TMZ-induced PTEN nuclear import depends on PTEN ubiquitylation modification by Smurf1. The Smurf1 suppression decreases the TMZ-induced PTEN nuclear translocation and enhances the DNA damage. In addition, Smurf1 degrades cytoplasmic PTEN K289E (the nuclear-import-deficient PTEN mutant) to activate the PI3K/Akt pathway under TMZ treatment. Altogether, Smurf1 interconnectedly promotes PTEN nuclear function (DNA repair) and cytoplasmic function (activation of PI3K/Akt pathway) to resist TMZ. These results provide a proof-of-concept demonstration for a potential strategy to overcome the TMZ resistance in PTEN wild-type GB patients by targeting Smurf1. | Smurf1 Suppression Enhances Temozolomide Chemosensitivity in Glioblastoma by Facilitating PTEN Nuclear Translocation
The tumor suppressor PTEN mainly inhibits the PI3K/Akt pathway in the cytoplasm and maintains DNA stability in the nucleus. The status of PTEN remains therapeutic effectiveness for chemoresistance of the DNA alkylating agent temozolomide (TMZ) in glioblastoma (GB). However, the underlying mechanisms of PTEN’s interconnected role in the cytoplasm and nucleus in TMZ resistance are still unclear. In this study, we report that TMZ-induced PTEN nuclear import depends on PTEN ubiquitylation modification by Smurf1. The Smurf1 suppression decreases the TMZ-induced PTEN nuclear translocation and enhances the DNA damage. In addition, Smurf1 degrades cytoplasmic PTEN K289E (the nuclear-import-deficient PTEN mutant) to activate the PI3K/Akt pathway under TMZ treatment. Altogether, Smurf1 interconnectedly promotes PTEN nuclear function (DNA repair) and cytoplasmic function (activation of PI3K/Akt pathway) to resist TMZ. These results provide a proof-of-concept demonstration for a potential strategy to overcome the TMZ resistance in PTEN wild-type GB patients by targeting Smurf1.
Glioblastoma (GB) is one of the most common malignant brain tumors, accounting for about 45% of primary malignant brain tumors and 15% of central nervous system tumors [1,2]. Despite extensive research efforts to better understand and treat these tumors, the prognosis for GB patients treated with Stupp standard procedures (surgery followed by radiation and chemotherapy) remains poor, with a median survival time fewer than 15 months [3]. Temozolomide (TMZ), which is more toxic in cancer cells with a slightly more alkaline pH than in normal cells, is the first-line chemotherapy agent in the treatment of GB [4]. However, the development of TMZ resistance often becomes the limiting factor in effective treatment. Thus, investigating mechanisms of TMZ resistance can help to identify novel drug targets and provide effective chemotherapies. Previous studies show that TMZ promotes the methylation at N7-guanine (the most common, 70%), followed by O6-guanine (6%, critical for tumor cytotoxic activity), and O3-Adenine (9%) [5]. The mismatch repair (MMR), which recognizes and removes the nucleotide mismatch (O6 methylguanine thymine) in the newly synthesized DNA chain, plays a vital role in the sensitivity of TMZ by the formation of accumulated DNA double-strand breaks, ultimately leading to cell apoptosis. Thus, the MMR deficiency causes TMZ resistance by mediating the formation of O6-methylated guanine (MG)-containing mismatches to facilitate the adaptive gene mutation [6]. There are other reasons accounting for TMZ resistance: (1) The ABC transporter from the blood–brain barrier (BBB) and glioma stem cells (GSCs) export foreign TMZ out of cells, leading to limited capacity to access the BBB (20% of its levels in the systemic circulation) [7,8]. (2) The high level of methylation of O6-methylguanine-DNA methyltransferase (MGMT) expression directly accounts for resistance by repairing the TMZ-induced O6-MG. The patients (approximately 60% of GB) with unmethylated MGMT have a lower survival rate [9,10]. The MGMT inhibitor O6-benzyl guanine upregulates the GB sensitivity to TMZ [5,11,12]. (3) Tumor DNA repair systems, such as the base excision repair (BER) and DNA strand break (DSB) repair, account for TMZ-induced alkylation and DSB failure [6,13,14,15,16]. In addition, within BER modulation, poly ADP ribose polymerase-1 (PARP-1) is essential for the recruitment of BER proteins and consequent DNA repair [13]. PARP-1 also promotes the activation of DSB repair proteins such as phosphorylated histone H2A.X, p53, and SMC1 by interaction with the DNA damage response kinase Ataxia telangiectasia mutated (ATM) [17]. The PARP-1 inhibitor overcomes TMZ resistance in MMR-deficient primary GB cells that are independent of BER [6]. A further study verifies that PARP-1 can also mediate PARylation of MGMT to TMZ resistance by repairing O6-MG DNA damage in GB [9]. The status of PTEN in GB (35% mutation rate) greatly influences TMZ resistance. By combining the EGFR tyrosine kinase inhibitor erlotinib with radiation therapy and TMZ, GB patients with MGMT gene silencing and intact PTEN have a significant survival advantage [18]. Studies show that the TMZ treatment is more effective in eradicating GB with PTEN loss [19], suggesting that PTEN is another molecular signature to affect GB patient survival. Mechanistically, cytoplasmic PTEN exerts its lipid phosphatase activity and dephosphorylates PIP3 to PIP2 in order to block PI3K pathways [20,21,22]. For example, the activation of the PI3K/Akt pathway increases the expression of PTEN, and the perturbations of PTEN status (mutation) limit the efficiency of PI3K inhibitors (BYL-719 and AZD8186) in tumors [23]. The combination of PI3Ki/PARPi presents an efficient therapeutic approach in PTEN-deficient tumors [24]. In addition, accumulating evidence has shown that nuclear PTEN is involved in the regulation of DNA damage repair, chromosome stability, and cell-cycle progression in its phosphatase-independent manner. The loss of PTEN leads to the accumulation of DSBs and genomic instability by impairing CHK1 function [25]. Nuclear PTEN promotes the tumor-suppressive activity of the APC-CDH1 complex for the fail-safe cellular senescence response [26]. The modification of PTEN is critical for cytoplasmic and nuclear localization and functions as a tumor suppressor. The existing studies show that post-translational modifications, such as mono-ubiquitination of PTEN by Nedd4 and de-ubiquitination by HAUSP/USP7, contribute to the nuclear import and export of PTEN, respectively [27,28,29]. Monoubiquitylation (K13 or K289), phosphorylation (S113), and SUMOylation (K254) promote PTEN nuclear translocation [30,31,32,33]. Zhang et al., mentioned that neddylation (K197 and K402) of PTEN regulates its nuclear import and promotes tumor development by dephosphorylation and stabilization of the fatty acid synthase [32]. It has been observed that the blockage of PTEN nuclear import promotes glioma sensitivity to chemo- or radiotherapy. For instance, the inhibition of nuclear PTEN by blocking phosphorylation of Y240-PTEN enhances the sensitivity to radiotherapy efficacy through attenuated DNA repair [34,35]. It also suggests the potential regulatory mechanisms that specifically regulate the abundance of nuclear PTEN in response to TMZ treatment. However, these mechanisms have rarely been studied. Elucidating this mechanism will help to learn the relationship between nuclear PTEN and TMZ drug resistance and provide a new perspective on the therapeutic depletion of nuclear PTEN in combination with TMZ. The HECT-type E3 ubiquitin ligase Smad ubiquitylation regulatory factor 1 (Smurf1) was initially identified to degrade SMADs through ubiquitination and the 26S proteasome in the TGF-β/BMP pathway [36]. Existing studies reported an inhibitory role of Smurf1 in cancer metastasis in lung cancer by targeting SRSF5 [37]. Other studies demonstrated that Smurf1 regulates multiple substrates, including UVRAG, Kindlin-2, ER-alpha, and p120-catenin, to promote tumor progression in different cancers [38,39,40,41]. Elevated Smurf1 in GB correlates with a worse prognosis [42]. We previously reported that Smurf1 suppression decreases the ubiquitination and degradation of PTEN to inhibit the PI3K/Akt pathway [43]. However, whether Smurf1 regulates PTEN nuclear import and drug resistance remains unclear. The obtained results show that TMZ promotes the nuclear import of PTEN dependent on Smurf1. Smurf1 knocking down decreases PTEN nuclear translocation, leading to enhanced TMZ-induced DNA damage. In summary, this study explores the therapeutic implications of Smurf1 and TMZ resistance, which may provide a promising treatment for GB.
GB cell lines LN229 (CRL-2611) and U87 (HTB-14) were purchased from American Type Culture Collection. GB cell lines U251 (HTX1725) and U343 (HTX2007) were purchased from Otwo Biotech (Shenzhen, China). Cells were transfected with lentiviral vector-harboring shScramble (5′-CCTAAGGTTAAGTCGCCCTCGCTCGAGCGAGGGCGACTTAACCTTAGG-3′) and Smurf1-shRNA (Open Biosystems) (5′-GCCCAGAGATACGAAAGAGAT-3′) using the VigoFect transfection reagent (Vigorous Biotechnology Beijing Co., Ltd.) according to the manufacturer’s instructions. After obtaining the lentivirus as described above, stable knockdown of endogenous human Smurf1 was achieved by lentivirus infection. All cells were cultured in DMEM (C11995500BT, Gibco) with fetal bovine serum (10%) and antibiotics penicillin (100 U/mL)/streptomycin (100 μg/mL) in an incubator (5% CO2, 37 °C). The TMZ-resistant glioma cell line (LN229R) was obtained by gradually exposing the parental cell line (LN229) to medium with increasing doses of TMZ (2–100 µM, 85622-93-1, MACKLIN) over 2 months. The medium containing TMZ was changed every 2–3 days. In addition, 250 μM TMZ was used for subsequent analysis. Primary MEFs were obtained from embryos at 14–16 days post coitum. The uterine horns of anesthetized pregnant mice were rinsed in ice-cold sterile PBS. The embryos were separated from their placenta and placed in a 60 mm plate containing ice-cold sterile PBS. The head, liver, and gut from the embryo were removed and the remaining portion put in a 60 mm plate containing trypsin EDTA. Then, they were cut into small pieces (~1 mm3) using a sterile razor blade and scissors. The chopped materials were incubated with 5 mL of trypsin EDTA for 30 min at 37 °C. Then, 1 mL DMEM (with FBS and penicillin/streptomycin) was added to the chopped embryos, followed by centrifuging at 1000 rpm speed for 5 min. The pellets were resuspended and transferred to a 60 mm dish containing DMEM (with FBS and penicillin/streptomycin). Cells were cultured in 12-well plates and grown for 12 h before treatments with TMZ. Cell numbers were recorded by TC10 Automated Cell Counter (Bio-Rad, Hercules, California, USA) at indicated days.
The cells were seeded at a density of 104 cells/well in 96-well plates. The cells were incubated at 37 °C, then 10 μL MTT (10 mg/mL, Solarbio, Beijing, China, M8180) was added to wells for 4 h at 37 °C. Finally, the 150 μL DMSO was added, and the absorbance was measured at the wavelength of 490 nm on the microplate reader to analyze cell proliferation and viability.
pCMV-PTEN was obtained from Addgene. A full-length PTEN cDNA was amplified from pCMV-PTEN, then was loaded onto 3×Flag vector. The point mutation of Flag-PTEN was generated by the site-directed mutagenesis with the following primers: 5′-TCAGAAGAAGTAGAAAATGGA-3′ and 5′-GGTTTCCTCTGGTCCTGGTAT-3′ for Flag-PTENK289E. The mutations were confirmed by sequence analysis. The human siRNAs were purchased from JTSBIO (Wuhan, China): si-Smurf1 was 5′-GCGUUUGGAUCUAUGCAAATT-3′, si-Smurf1-1 was 5′-CCAGGGAGUGGCUUUACUUTT-3′, si-PTEN was 5′-CCACCACAGCUAGAACUUATT-3′, and si-PTEN-2 was 5′-GGTGTAATGATATGTGCAT-3′. The cells were transfected with plasmids or siRNAs using Lipofectamine 2000 (Invitrogen, Waltham, MA, USA) or RNAiMAX reagent (Invitrogen, Waltham, MA, USA) following the supplier’s instructions, respectively. The overexpression or knockdown efficiency was examined by Western blotting.
Mice were housed in specific pathogen-free facilities, and the Ethics Review Committee for Animal Experimentation of the Beijing Institute of Technology University approved the experimental protocol. Smurf1WT and Smurf1KO mice were a kind gift from Dr. Lingqiang Zhang (Beijing Institute of Lifeomics, China). The cells were inoculated subcutaneously into Male BALB/c nude mice (6–7 weeks old; SPF (Beijing) Biotechnology Co., Ltd.). The length (L) and width (W) of the tumor were measured with a caliper every 7 days, and tumor volumes were calculated using the equation volume = (π × L × W2)/6. For TMZ treatment, mice were intraperitoneally injected with 20 mg/kg TMZ every 2 days for 5 weeks. The mice were anesthetized and euthanized, and the tumors were removed, imaged, and weighed. The Beijing Institute of Technology University Institutional Animal Care and Use Committee approved all animal studies in this study. The study is compliant with all relevant ethical regulations involving the manipulation of experimental animals.
The RIPA buffer (50 mM pH 7.6 Tris–HCl, 150 mM NaCl, 0.5% sodium deoxycholate, 1% NP-40, and protease inhibitor cocktail from Roche) with PMSF and phosphatase inhibitor was used to lyse cells. The samples were separated via SDS-PAGE and transferred to nitrocellulose filter membrane. The nitrocellulose filter membrane was blocked by 5% skim milk, followed by incubating with primary antibodies (12 h) and peroxidase-conjugated secondary antibodies (1 h). Finally, the bands were visualized by chemiluminescence reagents. Primary antibodies: β-actin (1:2000, Sigma, Taufkirchen, Germany, A1978-200), Smurf1 (1:500, Santa Cruz, Dallas, TX, USA, sc100616), p-p70 S6 kinase (Thr389) (1:500, Cell Signaling Technology, Danvers, MA, USA, #9205), p70 S6 kinase (1:500, Santa Cruz, Dallas, TX, USA, sc-8418), p-Akt (Ser473) (1:1000, Cell Signaling Technology, Danvers, MA, USA, #4060), Akt (1:1000, Cell Signaling Technology, Danvers, MA, USA, #9272), p-mTOR (Ser2448) (1:500, Cell Signaling Technology, Danvers, MA, USA, #2971), mTOR (1:1000, Cell Signaling Technology, Danvers, MA, USA, #2983), PTEN (1:500, Santa Cruz, Dallas, TX, USA, sc-7974), Flag (1:1000, Sigma, Taufkirchen, Germany, F3165), PARP (1:500, Santa Cruz, Dallas, TX, USA, sc-8007), Ub (1:1000, Abclonal, Wuhan, China, A19686), and γH2A.X (Ser139) (1:1000, Abcam, Cambridge, UK, ab26350). Secondary antibodies were used at 1:5000 dilution: Horseradish Peroxidase conjugated goat anti-rabbit IgG (BOSTER, Pleasanton, CA, USA, BA1054), Horseradish Peroxidase conjugated goat anti-mouse IgG (BOSTER, Pleasanton, CA, USA, BA1050). Relative protein levels were quantified by scanning densitometry, and the relative gray value of proteins corrected for background was calculated as: (band intensity of protein of interest)/(band intensity of loading control).
The cells were cultured on coverslips and fixed with 4% paraformaldehyde for 15 min and washed twice with PBS. The slides were treated with 0.1% TritonX-100 (PBS) for 5 min and washed twice with PBS, then blocked with 5% BSA for 1 h. Subsequently, the slides were incubated by primary antibody Flag (1:500, Sigma, Taufkirchen, Germany, F3165), PTEN (1:200, Santa Cruz, Dallas, TX, USA, sc-7974), or γH2A.X (Ser139) (1:1000, Abcam, Cambridge, UK, ab11174) for 12 h and fluorescently labeled by secondary antibody Alexa Fluor® 488 goat anti-mouse IgG (Life Technologies, Waltham, MA, USA, A11001) for 1 h. The cell nuclear was stained by Fluor shield mounting medium with DAPI (Abcam, Cambridge, UK). The samples were observed on Nikon N-SIM microscope.
Cells were collected and lysed with sucrose buffer with 0.5% NP40 for 40 min on ice (sucrose buffer: 0.32 M sucrose, 3 mM CaCl2, 2 mM MgAc, 0.1 M EDTA, 1 mM DTT, 0.5 mM PMSF). The lysate was separated via centrifuging at 600× g for 15 min at 4 °C. The supernatant was collected as cytoplasm. The precipitate was resuspended with sucrose buffer and centrifuged at 600× g for 15 min at 4 °C 4 times. The precipitate was resuspended with RIPA buffer with PMSF and phosphatase inhibitors as the nucleus.
The sample was centrifuged at 12,000 rpm for 10 min at 4 °C. The rProtein G beads (Solarbio, Beijing, China, R8300) were washed with PBS at 13,500× g for 2 min at 4 °C 3 times. The anti-PTEN (Santa Cruz, Dallas, TX, USA, sc-7974) was added to beads and incubated for 4 h at 4 °C. Then, the antibody was removed, and the beads were washed with PBS at 3000× g for 2 min at 4 °C 6 times. The sample was incubated with the beads for 12 h at 4 °C. After the incubation, the beads were washed with 1 × PBS 4 times. Both samples were identified by Western blotting.
We previously reported PI3K/Akt signaling is prohibited in LN229 (PTEN wild type), compared to U343 (PTEN wild type) and U87/U251 (PTEN deficiency) [42]. A previous report showed that the upregulated PI3K/Akt pathway contributes to TMZ resistance [44]. It reported that RNA modification regulator adenosine deaminases acting on RNA (ADARs) promote TMZ resistance by upregulating p-Akt. U343 expressing low ADARB1 may restore TMZ sensitivity in GB cells by decreased p-Akt [45]. To study whether the status of PTEN and/or Akt signaling affects the sensitivity of TMZ, we first analyzed the publicly available dataset in the Genomics of Drug Sensitivity in Cancer (GDSC) and The Cancer Genome Atlas Program (TCGA-GB) where GB samples were classified into two groups according to status of PTEN: wild type (n = 16 in GDSC-GB and n = 102 in TCGA-GB) and mutant (n = 17 in GDSC-GB and n = 47 in TCGA-GB). Results support the idea that there is no statistical difference in TMZ IC50 between the PTEN WT and PTEN mutant group (Figure 1a,b). Specifically, the PTENLow subgroup (n = 12 in TCGA-GB) had significantly worse survival compared with the PTENHigh subgroup (n = 17 in TCGA-GB) in PTEN wild-type patients with TMZ treatment (Figure 1c). Thus, we propose that PTEN wild-type expression made a difference on TMZ resistance. Additionally, the IC50 for TMZ of the U343 cell line was not included in the GDSC dataset. We then tested the cytotoxic effect of TMZ in the LN229, U251, U343, and U87 cell lines. The IC50 for TMZ of cell lines was as follows: LN229, 450 ± 50 μM; U87, 285 ± 40 μM; U251, 269 ± 40 μM; U343, 31 ± 10 μM (Figure A1a,b). These data showed that LN229, but not U343, was more resistant to TMZ treatment than the PTEN-deficient U87 and U251 cell lines. We also treated LN229, U343, U87, and U251 cell lines with TMZ, finding that LN229, but not U343, was more resistant to TMZ treatment than the PTEN-deficient U87 and U251 cell lines (Figure 1d). We then assumed that TMZ feedback induces the activation of PI3K/Akt signaling. Indeed, we found that TMZ hyperactivated the PI3K/Akt pathway in LN229, U251, and U87 cells, evidenced by upregulation of p-AktS473, p-mTORS2448, and p-p70S6KT389, but not U343 (Figure 1e). These results suggest the inability to activate the PI3K/Akt pathway may be associated with U343 sensitivity to TMZ. Smurf1 is overexpressed in GB cells and promotes cell growth. Given that our previous report showed that Smurf1 suppression decreases the ubiquitination and degradation of PTEN to inhibit the PI3K/Akt pathway [43], we analyzed the overall survival in the Smurf1 high group (n = 29, including 24 patients (n = 23 with wild-type PTEN) with TMZ treatment) and Smurf1 low group (n = 30, including 28 patients (n = 20 with wild-type PTEN) with TMZ treatment) in the TCGA GB patient dataset, and found that the Smurf1 high group had significantly worse survival compared with low group (Figure 1f). To evaluate whether suppression of Smurf1 could restore cell growth and sensitize the cells to TMZ treatment, we transfected GB cells with si-Smurf1 and si-Control RNA and treated with or without TMZ. Consistently, we found no statistical difference in the efficacy of treatment of PTEN-deficient U251 and U87 cells between the dual-treatment group (si-Smurf1 + TMZ) and the TMZ alone group (Figure 1g). Importantly, the dual-treatment group showed reduced cell survival compared to the TMZ alone group in both LN229 and U343 cells (Figure 1g). Further, PTEN knockdown reduced and PTEN overexpression restored the effect of dual treatment in LN229 and U251, respectively, implying that Smurf1 promotes TMZ resistance in a PTEN-dependent manner (Figure A1c,d). We then knocked down Smurf1 in LN229 and U251 to investigate whether suppression of Smurf1 sensitizes TMZ treatment by inhibiting the PI3K/Akt pathway. We found that the activation of the PI3K/Akt pathway is inhibited in the dual-treatment compared with the TMZ alone treatment in LN229, but not in U251 (Figure 1h,i). Of note, PTEN knockdown inhibited si-Smurf1-reduced suppression of the PI3K/Akt pathway in LN229 (Figure A1e–g), suggesting that suppression of Smurf1 inhibits Akt signaling in a PTEN-dependent manner. We next investigated whether Smurf1-mediated hyperactivation of PI3K/Akt depends on nuclear PTEN. We transfected Flag-PTENK289E, which is defective in nuclear import [46], into U251 cells and found that Smurf1 knockdown still decreased p-AktS473, p-mTORS2448, and p-p70S6KT389, demonstrating that Smurf1 degrades cytoplasmic PTEN to activate PI3K/Akt under TMZ treatment (Figure 1j), and the inhibition of Smurf1 to prohibit Akt signaling is nuclear PTEN independent. More importantly, Smurf1 knockdown combined with TMZ treatment significantly decreased the cell proliferation of U251 with Flag-PTENK289E (Figure 1k). These results illustrate that Smurf1-mediated hyperactivation of PI3K/Akt is sufficient to induce TMZ resistance independent of nuclear PTEN.
PARP-induced phosphorylated H2A.X (γH2A.X) is one of the DNA damage markers by instant accumulation at nascent DSB sites [13,47]. Interestingly, we found the dual treatment led to an increase in PARP and γH2A.X in LN229, while the TMZ alone group and treated U251 cells showed no similar increase (Figure 2a–d and Figure A2a,b), suggesting suppression of Smurf1 may also promote DNA damage. Of note, transfection of PTEN in U251 had no effect on PARP and γH2A.X expression between the si-Smurf1 and si-Control group (Figure A2c,d), suggesting that TMZ is the inducer of the DNA damage. To further confirm the protective role of nuclear PTEN in TMZ-induced DNA damage, we transfected Flag-PTEN or Flag in U251, and si-PTEN or si-Control RNA in LN229, respectively. By overexpression of PTEN in U251, Smurf1 suppression rescued the promotion effect by increased γH2A.X and PARP protein levels in response to TMZ (Figure 2e,f). Conversely, the combination of si-Smurf1 and TMZ failed to induce the synthetic effect in LN229 with knocking down of PTEN (Figure 2g,h and Figure A2e–h). These results imply that Smurf1-suppression-enhanced TMZ-induced DNA damage is due to the nuclear translocation of PTEN. To investigate whether Smurf1 increased the nuclear translocation of PTEN under TMZ treatment, we performed a cytoplasmic and nuclear protein extraction experiment in LN229 and U251-Flag-PTEN cells. TMZ treatment increased the nuclear fraction of PTEN and decreased the cytoplasmic proportion of PTEN in LN229 under TMZ treatment (Figure 2i–l), suggesting that TMZ induces PTEN nuclear translocation. Notably, Smurf1 knockdown significantly inhibited PTEN nuclear translocation in LN229 and U251-Flag-PTEN in TMZ treatment (Figure 2i–l), indicating that TMZ-induced PTEN nuclear translocation is dependent on Smurf1. To quantify the percentage of PTEN+ cells with or without TMZ treatment in the si-Smurf1 or si-Control groups, we separated PTEN+ cells into three groups: (1) nuclear PTEN+, cytoplasmic PTEN is less than 35% nuclear PTEN immunofluorescent intensity (cytoplasm < nuclear), (2) 65% nuclear > cytoplasm > 35% nuclear, (3) cytoplasmic PTEN+, cytoplasm > 65% nuclear. We found that the percentage of nuclear PTEN+ cells (around 60%) was significantly decreased by suppression of Smurf1 under TMZ treatment in U251-Flag-PTEN, U343, and LN229 cells (Figure A2i–m), which further demonstrates that Smurf1 promotes nuclear PTEN import. Next, we studied whether PTEN nuclear import depends on PTEN ubiquitylation modification by Smurf1. LN229 cells were treated with TMZ, and immunoprecipitation was performed using a PTEN antibody. The nuclear fraction was subjected to SDS-PAGE and probed with the Ub antibody. TMZ treatment significantly increased the Ub level of nuclear PTEN (Figure 2m). Furthermore, Smurf1 knockdown significantly reduced the ubiquitination level of nuclear PTEN in LN229 with TMZ treatment (Figure 2m), suggesting that Smurf1 mediates PTEN modification by its ubiquitylation. These data indicate that Smurf1 suppression downregulates TMZ-induced nuclear import of PTEN by inhibiting its ubiquitination. It has been shown that the enhanced nuclear PTEN accumulation is associated with increased mono-ubiquitination at the K289 site [46]. To investigate whether Smurf1 promotes TMZ-induced PTEN nuclear import by K289 mono-ubiquitylation, U251 cells were transfected with Flag-PTENK289E followed by TMZ treatment. Consistently, TMZ promotes PTEN wild-type nuclear import, but not PTENK289E, suggesting that TMZ promoted PTEN nuclear import by mono-ubiquitination at the K289 site (Figure A2n). The cytoplasmic and nuclear protein extraction of PTENK289E expressing U251 with TMZ treatment showed that Smurf1 knockdown has no significant difference in the protein level of γH2A.X, indicating that ubiquitination of K289 PTEN plays an essential role in Smurf1-mediated DNA repair (Figure 2n,o). These results implied that Smurf1 facilitates PTEN nuclear translocation to repair DNA damage under TMZ treatment by K289 mono-ubiquitination.
To investigate the efficacy of Smurf1 suppression in TMZ-resistant GB cell lines, we cultured LN229R and U343R cells, which are TMZ resistant, by exposing LN229 or U343 cells to stepwise increasing concentrations of TMZ (2–100 µM) over 2 months. The cell viability assay showed that LN229R or U343R had significantly higher TMZ tolerance than LN229 or U343 cells (Figure 3a and Figure A3a). Next, we knocked down Smurf1 in LN229 and LN229R under TMZ treatment. The results showed that the dual-treatment group decreased cell survival compared to the single TMZ treatment in LN229R, implying that Smurf1 suppression inhibited the proliferation of TMZ resistance cells (Figure 3b). Additionally, LN229R showed an increased significant PI3K/Akt pathway activation compared with parental LN229 cell lines, evidenced by upregulation of p-AktS473, p-mTORS2448, and p-p70S6KT389 (Figure 3c). To test the underlying effectiveness of si-Smurf1 in TMZ re-sensitivity by enhanced DNA damage and suppression of Akt signaling, we transfected LN229R and U343R with si-Smurf1. We found that dual treatment of siSmurf1 and TMZ increased γH2A.X and reduced PI3K/Akt (p-AktS473, p-mTORS2448, and p-p70S6KT389) compared with the TMZ treatment alone (Figure 3d–f and Figure A3b–e). Importantly, it also showed that PTEN knockdown diminished the effectiveness of si-Smurf1 in LN229R and U343R, suggesting LN229R and U343R were re-sensitive to TMZ by targeting Smurf1 in a PTEN-dependent manner (Figure 3d–g and Figure A3f–i). Taken together, Smurf1 interconnectedly regulates PTEN both to activate PI3K/Akt (degradation of cytoplasmic PTEN) and repair DNA damage (import of nuclear PTEN) responding to TMZ treatment.
To examine the combined effect of Smurf1 suppression and TMZ in vivo, mouse embryonic fibroblasts (MEFs) were generated from Smurf1+/+ mice and Smurf1−/− mice. Consistently, TMZ-treated Smurf1−/− MEFs displayed a decreased PI3K/Akt pathway (downregulation of p-AktS473, p-mTORS2448, p-p70S6KT389) and induced DNA damage (upregulation of PARP and γH2A.X) compared to TMZ-treated WT MEFs (Figure 4a–c). To further explore the synthetically lethal effect, the LN229-shSmurf1 and LN229-shScramble cells were subcutaneously injected into the left and right flanks of eight-week-old female nude mice. Fourteen days after implantation, the mice were treated intraperitoneally with TMZ or normal saline every two days for five weeks (Figure 4d, e). Significantly, TMZ suppressed the growth rate, size, and weight of tumors in LN229-shSmurf1 compared to in LN229-shScramble, suggesting that the combination of Smurf1 inhibition and TMZ treatment synthetically inhibited tumor growth (Figure 4f–h). In sum, Smurf1 interconnectedly regulates PTEN both to activate PI3K/Akt (degradation of cytoplasmic PTEN) and repair DNA damage (import of nuclear PTEN) to resist TMZ. The co-treatment of Smurf1 suppression and TMZ has potential efficacy for PTEN wild-type GB (Figure 5).
Clinical findings suggest that PTEN gene alterations are associated with poor prognosis and may influence tumor responses to current therapies in high-grade GB [48]. Previous reports show only 25% of cancer patients showing a correlation between loss of PTEN and loss of its mRNA [49,50]. Our result indicates that understanding the protein levels and localization of PTEN in patient tumors may further help stratify these patient populations. According to a relevant survey, 7.1% of GB patients manifested nuclear positivity for PTEN [51]. Several studies reported that nuclear PTEN plays a crucial role in regulating DNA damage repair [34]. However, the role of nuclear PTEN in response to chemotherapy resistance remains largely unknown. Recent studies demonstrated that ionizing radiation (IR) induces drug resistance by promoting nuclear PTEN import independent of its lipid phosphatase activity [34]. This study revealed that Smurf1 contributes to TMZ chemoresistance by promoting PTEN nuclear import. PTEN nuclear translocation is a dynamic process [52]. Ubiquitination, SUMOylation, and phosphorylation have been shown to contribute to the translocation of PTEN. A previous study has proven that E3 ligase Nedd4-1, WWP1, WWP2, and Smurf1 target cytoplasmic PTEN for ubiquitin-mediated proteasome degradation [43,53,54,55]. Monoubiquitylation of PTEN by Nedd4-1 also contributes to PTEN nuclear import [46,52]. Therefore, this study aims at studying whether Smurf1 selectively induced ubiquitylation of PTEN to promote PTEN nuclear import. Here, we showed that Smurf1 knockdown reduces nuclear PTEN import in U251 with PTEN overexpression. In addition, Smurf1 knockdown decreases the protein level and ubiquitylation of nuclear PTEN in response to TMZ in PTEN wild-type LN229, demonstrating that Smurf1 promotes ubiquitylation of PTEN and its nuclear import. Furthermore, we identified that TMZ promotes PTEN nuclear import by mono-ubiquitination at the K289 site, which may be mediated by Smurf1. Therefore, what is the role of the imported nuclear PTEN in response to TMZ? Previous studies have shown that nuclear PTEN is involved in maintaining genome stability by promoting DNA repair. They also demonstrated that the modifications of PTEN in the nucleus play a controversial role in tumor development. The phosphorylation of PTEN at Y240 sites promotes the recruitment of RAD51 to accelerate DNA repair by facilitating its interaction with Ki-67 [34]. A SUMOylation of PTEN at K254 is required for homologous recombination repair of DSBs through the DNA-damage-induced protein kinase ATM cascade [33]. However, the neddylation of PTEN at K197 and K402 sites by XIAP promotes tumorigenesis and progression through its interaction with FASN, in order to increase de novo fatty acid synthesis [32]. In this case, we proposed that the Smurf1-induced PTEN nuclear translocation to facilitate resistance for TMZ may also be mediated by chromosomal stability and DNA repair capacity. Indeed, Smurf1-mediated PTEN nuclear translocation protects DNA against being damaged, which is indicated by the decreased DNA DSBs marker γH2A.X. Notably, PARP-1/2 inhibitor ABT-888 enhances the TMZ efficacy in PTEN-deficient GB [48]. Consistently, similar to PTEN-deficient lines (U251 and U87), the PTEN wild type line (LN229) with Smurf1 knockdown also has high protein level of DNA damage signaling protein PAPR-1 and γH2A.X in response to TMZ. Accumulating evidence shows that the PI3K/Akt/mTOR pathway is activated in response to TMZ, which is another essential factor in drug resistance [56,57]. In addition to their critical roles in balancing anabolic and catabolic responses, Akt and its downstream effector mTORC1 react to changes in genomic integrity [58,59]. Akt has been implicated in regulating diverse substrates that have nuclear functions, including important roles in DNA damage response (DDR), DNA repair, and the maintenance of genomic stability [60]. The study demonstrates that DSB promotes accumulation and co-localization of p-AktS473 with γH2AX and p-ATMS1981, which makes non-homologous end-joining-mediated DSB repair start [61,62]. Activated Akt in the nucleus (immediately after Ser473 phosphorylation by DNAPK) conceivably regulates phosphorylation of the transcription factor FOXO, which resides in the nucleus prior to Akt phosphorylation [63,64], or DNA repair effector proteins, such as XLF or MERIT40 [59,65]. Moreover, Akt shuttles between nuclear and cytoplasmic compartments and thereby carries a nuclear signal (via phosphorylation) back out to the cytoplasm (inside-out signaling) to phosphorylate its downstream targets. Inhibition of mTORC1 signaling enhances DNA damage sensitivity [66]. Thus, DNA-damage-induced Akt activation expedites crosstalk between DNA repair/genome stability and cell growth/survival pathways [67]. In addition, inhibiting cellular DNA repair factors block Akt activation and promote reactivation. Whether PTEN affects Akt activity by repairing DNA damage deserves further confirmation. The interaction between nuclear PTEN (as DNA repair factor) and Akt requires further study. Similar to Nedd4-1 and WWP1, Smurf1 can ubiquitylate and degrade cytoplasmic PTEN, thus positively regulating PI3K/Akt [43,53,55,68]. Decreased cytoplasmic PTEN confers TMZ resistance by promoting proliferation in TMZ-resistant cells. How does Smurf1 promote TMZ resistance by synthetical regulation of nuclear and cytoplasmic PTEN? This study assumes that Smurf1 can ubiquitylate PTEN for both cytoplasmic degradation and nuclear import in response to TMZ. The phosphatase-inactivating mutation of PTEN, such as G129R and C124S, enhances the ability of PTEN nuclear import in response to IR [34]. Another study also confirmed that PTEN knockdown increases nuclear p-Akt [69], which may indicate that the reduction in nuclear PTEN increases DNA repair and leads to p-Akt nuclear transport. However, the interconnected determination of nuclear import by mono- or poly-ubiquitylation modification is still unclear. Using PTENK289E mutation to inhibit the mono-ubiquitylation of PTEN nuclear import, it is deduced that Smurf1 still degrades cytoplasmic PTENK289E protein levels to activate the PI3K/Akt signaling pathway under TMZ treatment. Smurf1-increased cytosolic PTEN leads to a decreased mTOR signal and further increases sensitivity to DNA damage. The present study suggests that Smurf1 interconnectedly regulates PTEN either to activate PI3K/Akt (degradation of cytoplasmic PTEN) or repair DNA damage (import of nuclear PTEN) to resist TMZ (Figure 4g). It is important to know whether targeting Smurf1 is effective in the TMZ resistance cell line. We also generated LN229-R cell lines and found that targeting Smurf1 could induce cell re-sensitivity to TMZ by protecting DNA damage and reducing the PI3K/Akt pathway in a PTEN-dependent manner. The MEFs from Smurf1+/+ mice and Smurf1−/− mice, as well as the mice model of subcutaneous GB, also confirmed the combination of Smurf1 inhibition and TMZ treatment further inhibited tumor growth. The combination of TMZ and Smurf1 inhibition may have a therapeutic effect on patients with PTEN wild-type GB. The dual PI3K/mTOR inhibitors, such as PI-103 and NVP-BEZ235, are synergistic with TMZ to increase the therapeutic efficiency [57]. In this study, the Smurf1 suppression can also inhibit the PI3K/Akt pathway. Moreover, phosphorylated Akt increases MGMT expression by reducing protein level and nuclear import transcription factor FKHRL1/FOXO3A [70]. Whether Smurf1 increases MGMT by activating Akt to resistant TMZ requires further investigation. Moreover, the miR-26a inhibitor synergistically decreases TMZ resistance by inhibiting the stemness of GSCs [71]. Smurf1 knockdown reduces the expression of stem cell markers (Sox2 and Nestin) by reactivating PTEN [43]. Whether Smurf1 inhibition regulates tumor stemness to re-sensitive TMZ requires further investigation. In addition, previous studies have shown that the PTEN deficiency synergistically works with PARP inhibitors by damaging DNA repair [72]. A further study warrants testing the combination of PARP inhibitor with TMZ in Smurf1 knockdown GB cells with the PTEN wild type. It is confirmed that the Smurf1 suppression not only inhibits the PI3K/Akt pathway but also increases the γH2A.X protein level in response to TMZ. Smurf1 is a novel target in TMZ resistance in PTEN wild-type cells, and it provides a new treatment scheme for TMZ re-sensitivity through multiple channels.
TMZ resistance is a major problem in the treatment of malignant brain tumors. However, the mechanisms underlying TMZ-resistant GB cells have not been fully characterized. Here, we report that TMZ-induced PTEN nuclear import depends on PTEN ubiquitylation by Smurf1. The Smurf1 suppression decreases the TMZ-induced PTEN nuclear translocation and enhances the DNA damage. In addition, Smurf1 degrades cytoplasmic PTENK289E (the nuclear-import-deficient PTEN mutant) in order to activate the PI3K/Akt pathway under TMZ treatment. Altogether, Smurf1 interconnectedly promotes both PTEN nuclear function (DNA repair) and cytoplasmic function (activation of PI3K/Akt pathway) to resist TMZ. These results provide a proof-of-concept demonstration for a potential strategy to treat PTEN wild-type GB patients who acquired TMZ resistance through targeting Smurf1. | true | true | true |
|
PMC9600605 | 36069441 | Jane X. Yeh,Yunfan Fan,Maggie L. Bartlett,Xiaoyan Zhang,Norah Sadowski,Debra A. Hauer,Winston Timp,Diane E. Griffin | Treatment of Sindbis Virus-Infected Neurons with Antibody to E2 Alters Synthesis of Complete and nsP1-Expressing Defective Viral RNAs | 07-09-2022 | alphavirus,defective viral genomes,nanopore sequencing,neuron infection,viral encephalitis,virus clearance | ABSTRACT Alphaviruses are positive-sense RNA viruses that are important causes of viral encephalomyelitis. Sindbis virus (SINV), the prototype alphavirus, preferentially infects neurons in mice and is a model system for studying mechanisms of viral clearance from the nervous system. Antibody specific to the SINV E2 glycoprotein plays an important role in SINV clearance, and this effect is reproduced in cultures of infected mature neurons. To determine how anti-E2 antibody affects SINV RNA synthesis, Oxford Nanopore Technologies direct long-read RNA sequencing was used to sequence viral RNAs following antibody treatment of infected neurons. Differentiated AP-7 rat olfactory neuronal cells, an in vitro model for mature neurons, were infected with SINV and treated with anti-E2 antibody. Whole-cell RNA lysates were collected for sequencing of poly(A)-selected RNA 24, 48, and 72 h after infection. Three primary species of viral RNA were produced: genomic, subgenomic, and defective viral genomes (DVGs) encoding the RNA capping protein nsP1. Antibody treatment resulted in overall lower production of SINV RNA, decreased synthesis of subgenomic RNA relative to genomic RNA, and suppressed production of the nsP1 DVG. The nsP1 DVG was packaged into virus particles and could be translated. Because antibody-treated cells released a higher proportion of virions with noncapped genomes and transient transfection to express the nsP1 DVG improved viral RNA capping in antibody-treated cells, we postulate that one mechanism by which antibody inhibits SINV replication in neurons is to suppress DVG synthesis and thus decrease production of infectious virions containing capped genomes. | Treatment of Sindbis Virus-Infected Neurons with Antibody to E2 Alters Synthesis of Complete and nsP1-Expressing Defective Viral RNAs
Alphaviruses are positive-sense RNA viruses that are important causes of viral encephalomyelitis. Sindbis virus (SINV), the prototype alphavirus, preferentially infects neurons in mice and is a model system for studying mechanisms of viral clearance from the nervous system. Antibody specific to the SINV E2 glycoprotein plays an important role in SINV clearance, and this effect is reproduced in cultures of infected mature neurons. To determine how anti-E2 antibody affects SINV RNA synthesis, Oxford Nanopore Technologies direct long-read RNA sequencing was used to sequence viral RNAs following antibody treatment of infected neurons. Differentiated AP-7 rat olfactory neuronal cells, an in vitro model for mature neurons, were infected with SINV and treated with anti-E2 antibody. Whole-cell RNA lysates were collected for sequencing of poly(A)-selected RNA 24, 48, and 72 h after infection. Three primary species of viral RNA were produced: genomic, subgenomic, and defective viral genomes (DVGs) encoding the RNA capping protein nsP1. Antibody treatment resulted in overall lower production of SINV RNA, decreased synthesis of subgenomic RNA relative to genomic RNA, and suppressed production of the nsP1 DVG. The nsP1 DVG was packaged into virus particles and could be translated. Because antibody-treated cells released a higher proportion of virions with noncapped genomes and transient transfection to express the nsP1 DVG improved viral RNA capping in antibody-treated cells, we postulate that one mechanism by which antibody inhibits SINV replication in neurons is to suppress DVG synthesis and thus decrease production of infectious virions containing capped genomes.
Alphaviruses (family Togaviridae) are important causes of viral encephalomyelitis worldwide and a significant public health concern due to their high mortality rate, potential for permanent neurologic damage, and limited availability of vaccines and therapeutics (1, 2). Alphaviruses are transmitted by mosquitoes and are expanding in worldwide geographic range due to climate change and globalization (3–6). Sindbis virus (SINV), the prototype alphavirus, has a positive-sense RNA genome of 11.7 kb that is 5′-capped, polyadenylated, and has two open reading frames: one encoding the nonstructural polyprotein that is subsequently processed into the individual nonstructural protein components (nsP1, nsP2, nsP3, and nsP4) and another encoding the structural polyprotein (capsid, E3, E1 and E2 envelope glycoproteins, 6K, and TF) (7). In the SINV life cycle, 3 primary viral RNA species are made: (i) the full-length positive-sense genomic RNA (gRNA) that is eventually incorporated into released infectious virions, (ii) the negative-sense genomic RNA that serves as a template for viral RNA replication, and (iii) the shorter positive-sense subgenomic RNA (sgRNA), which includes the latter third of the genome encoding the structural proteins (Fig. 1). The relative production of these 3 viral RNAs changes over the course of the viral life cycle. Upon infection, the positive-sense capped SINV RNA genome from the infecting virion is directly translated by host machinery to produce the nonstructural polyprotein P123 or P1234, depending on read-through of an opal stop codon between the nsP3 and nsP4 genes (10% occurrence) (8). nsP1 within the polyprotein anchors the replication complex to the plasma membrane and with nsP3 induces formation of the spherules in which RNA synthesis occurs (9). nsP4, the viral RNA-directed RNA polymerase (RdRp), is cleaved from P1234 and with P123 initiates synthesis of the negative-sense gRNA that serves as a template for production of more positive-sense viral RNAs. Initiation at the 5′ promoter of the negative-sense gRNA generates full-length positive-sense SINV gRNA, while initiation at the internal subgenomic promoter generates a shorter sgRNA that encodes the SINV structural proteins. Efficient transcription from the subgenomic promoter results in production of approximately 3-fold more sgRNA than gRNA, allowing for synthesis of more structural proteins that are incorporated into newly formed infectious virions (10). As the life cycle progresses, P123 is further cleaved by nsP2 into the individual nonstructural proteins nsP1, nsP2, and nsP3, and synthesis of negative-sense gRNA ceases while positive-sense RNA production continues (7). In humans, SINV infection typically results in rash, fever, arthritis, and myalgia with persistent musculoskeletal pain (1). In rodents, SINV infection results in a well-characterized encephalomyelitis that parallels disease observed with alphavirus-induced encephalomyelitis in humans (2). Neurons of the brain and spinal cord are the primary targets of infection, with the olfactory tract, hippocampus, brainstem, and spinal cord motor neurons being especially susceptible (3, 4). SINV infection in rodents provides a model for investigation of alphaviral encephalomyelitis pathogenesis and recovery. Following intranasal or intracranial infection with TE, a strain of SINV that does not cause fatal encephalomyelitis, weanling mice develop transient kyphoscoliosis, hind-limb paralysis, and muscular atrophy but recover in 2 to 3 weeks without lasting clinical or histological evidence of neuronal damage (5–8). Peak infectious viral titers in the brain and spinal cord are reached 3 to 5 days after infection, after which viral clearance begins (9). Infectious virus is no longer detectable in the central nervous system (CNS) by 7 to 8 days after infection; however, viral RNA levels decline much more slowly over weeks and continue to be detectable at low levels (10). This difference suggests that distinct mechanisms mediate clearance of infectious virus versus clearance of viral RNA from infected neurons, and both processes are essential to prevent continued or reactivated virus production. Studies of SINV infection of immunodeficient mice have revealed many aspects of the immune responses involved in recovery. In particular, antibody against the SINV E2 glycoprotein is an important component of the immune response. Severe combined immunodeficient (SCID) mice that lack functional adaptive immune responses are unable to clear SINV from the CNS and continue to produce infectious virus (11). Adoptive transfer of monoclonal anti-E2 monoclonal antibody (MAb) alone is sufficient to induce clearance of infectious virus within 2 days after transfer (11). In vitro studies with primary rat dorsal root ganglion (DRG) neurons (11–13) and immortalized neuronal cell lines (14, 15) further revealed that antibody must be bivalent to be efficacious and does not require interaction with other cell types. Effects of antibody treatment of infected neuronal cells in vitro include decreased viral translation/transcription, inhibition of viral budding, restoration of host translation/transcription, restoration of host ability to respond to type I IFN, restoration of baseline intracellular Na+/K+ cation concentrations, and increased host cell survival (12, 14–16). In differentiated rat AP-7 olfactory neuronal (dAP-7) cells (17, 18), antibody-mediated clearance of SINV is associated with initiation of a signaling cascade consisting of transient activation of NF-κB, induction of leukemia inhibitory factor (LIF) cytokine production, and subsequent prolonged activation of STAT3 that results in decreased synthesis of viral structural proteins and RNA and improved cell survival (16). Live-cell imaging of dAP-7 cells infected with recombinant SINV with Broccoli aptamer-tagged viral RNA also showed that anti-E2 antibody decreased overall viral RNA and improved cell survival (19). However, these studies were not able to determine the effects of antibody on the different viral RNAs. Using high-throughput direct Oxford Nanopore MinION viral RNA sequencing, we evaluated the effect of anti-E2 antibody on RNA in SINV-infected dAP-7 olfactory sensory neuronal cells and found that antibody treatment induces changes in RNA abundance and type. Anti-E2 treatment improved production of rat RNAs and decreased production of SINV RNAs relative to untreated virus-infected cells. Antibody further affected the types of SINV RNAs produced, specifically decreasing sgRNA synthesis relative to gRNA synthesis and suppressing production of two previously unrecognized SINV RNAs containing the 5′ end, the nsP1 gene, and the 3′ untranscribed region (UTR) characteristic of deletion-type defective viral genomes (DVGs) (Fig. 1). Antibody treatment decreased the proportion of virions containing capped infectious gRNAs versus noncapped noninfectious gRNAs. The nsp1 DVGs could be translated and packaged, and we postulate these nsP1-containing DVGs have a proviral function through improved efficiency of SINV genome capping.
To evaluate changes in SINV RNA at the nucleotide level following anti-E2 antibody treatment of infected dAP-7 cells, nanopore sequencing was employed. Nanopore sequencing directly sequences RNA to avoid potential biases from either PCR or reverse transcription and provides long reads for read-through of full-length viral genomic RNA (11.7 kb). AP-7 cells were differentiated for 7 days into their nondividing form that models mature neurons (dAP-7 cells) (18, 20, 21) and then infected with the TE strain of SINV at a multiplicity of infection (MOI) of 10 based on plaques produced in BHK-21 cells, which results in initial infection of ~10% of the monolayer (18). Four hours after infection to allow for viral binding and entry, the cells were washed and treated with 5 μg/mL of monoclonal anti-E2 antibody, an amount within the range of antibody concentrations in brain (22), or medium alone as a negative control (untreated). At 24, 48, and 72 h after infection, total cellular RNA was collected and enriched for polyadenylated RNAs by oligo(dT)-conjugated magnetic bead sorting. The enriched mRNAs were then sequenced using a standard direct RNA nanopore method (Materials and Methods). Sequencing reads were aligned to the rat and SINV genomes to assess host and viral gene expression. By 24 h after infection, 76% of the total RNA reads were aligned with the SINV genome in untreated virus-infected cells (Fig. 2A), reflecting the high efficiency with which SINV inhibits host transcription. By 48 and 72 h after infection, the percentages of SINV reads increased to >80% and >85% of the RNA sequenced, respectively. Similar to previous studies that suggested that anti-E2 antibody treatment reduces the effect of SINV-induced transcription inhibition (15), anti-E2 antibody treatment resulted in a greater proportion of rat reads and lower proportion of SINV reads at all three time points assessed (Fig. 2A). At 24 h, 44% of the RNA in E2 antibody-treated cells aligned to the SINV genome. At 48 h and 72 h, these proportions stabilized at 49% and 41%, respectively. Within the SINV-aligned reads, three primary SINV RNA forms were observed: (i) the 11.7-kb positive-sense full-length SINV gRNA (nucleotides [nt] 1 to 11703), (ii) the shorter 3.7-kb SINV sgRNA (nt 7646 to 11703), and (iii) a truncated RNA containing the nsP1 gene (nt ~70 to 1700), the 3′ UTR, and sometimes the 5′ UTR of the SINV genome (Fig. 2B). These truncated RNAs resemble deletion-type defective viral genomes (DVGs) in structure and are subsequently subcategorized as “DVGs” (with end between nt 1749 and 1755 of the SINV genome) or “subDVG” (with end at or before nt 1745) depending upon the viral RNA end. The negative-sense gRNA was not detected as it is not polyadenylated and would not be captured with the magnetic bead pulldown. To evaluate changes in relative production of the different SINV RNA species with antibody treatment, we calculated the sequencing depth coverage across the SINV genome reference normalized to the total number of SINV reads in each run for each time point (Fig. 2B). We observed the expected 3′ bias due to the 3′→5′ directionality of nanopore sequencing as anchored to the poly(A) tail (23). Overall, coverage was greatest for the structural genes (capsid to E1) because this region is included in both the full-length SINV gRNA as well as the shorter sgRNA. Baseline levels of full-length SINV gRNA were reflected in the nonstructural gene region (nsP2 to nsP4) coverage. Coverage over both regions was relatively stable, consistent with synthesis of the SINV subgenomic and genomic RNAs as single molecules. The peaks in coverage at nsP1, the 3′ UTR, and the 5′ UTR reflect the separate truncated nsP1-containing DVGs identified. To determine how the SINV RNAs were individually affected by antibody treatment, a viral RNA abundance ratio was calculated by taking the coverage differences at each junction site and dividing by the total of these differences (Fig. 2C). In general, gRNA levels were inversely correlated with sgRNA levels across the samples. Comparing the antibody-treated versus untreated cells, the relative abundance of gRNA was higher at 24 h and 72 h, but abundances were comparable at 48 h after infection; however, the differences were not statistically significant due to variation across the triplicate samples. By the same token, the sgRNA relative abundances were lower at the 24-h and 72-h time points with antibody treatment but higher at 48 h. These data suggest that anti-E2 antibody, in addition to decreasing overall viral RNA levels, also differentially affects the types of SINV RNA produced. Association of anti-E2 treatment with higher proportions of gRNA and lower sgRNA could be explained by a delay in the transition from gRNA to predominantly sgRNA production that occurs during late infection and would be a valuable topic of future investigation. Antibody treatment also had a pronounced effect upon levels of the nsP1-containing DVGs. The antibody-treated cells had significantly lower DVG levels at all time points assessed (24 h, P < 0.0001; 48 h, P < 0.001; 72 h, P < 0.0001) and lower subDVG levels at 24 h (P < 0.01) than the untreated cells. These data suggest that in addition to antiviral antibody affecting the relative proportions of different SINV RNAs present, production of the nsP1-containing RNAs was especially sensitive to the effects of antibody.
The nanopore sequencing data showed that anti-E2 antibody treatment decreased levels of overall SINV RNA and affected the relative abundances of different SINV RNA species. To determine whether the antibody-induced changes reflect differences in RNA synthesis versus degradation and further investigate how shifts in viral RNA production are affected by antibody on a shorter time scale, dAP-7 cells were again infected with SINV (BHK-21 cell MOI of 10) and treated with medium (untreated) or 5 μg/mL anti-E2 antibody 4 h after infection (Fig. 3). At 0, 12, 24, 36, and 48 h after infection, the cells were pulsed with [3H]uridine for 2 h in the presence of dactinomycin to inhibit cellular transcription and allow for specific evaluation of viral RNA transcription. Whole-cell RNA was collected, separated by agarose-formaldehyde gel electrophoresis, and visualized by autoradiography. Two distinct bands corresponding to the SINV gRNA and sgRNAs were detected (Fig. 3A) and quantified by densitometry (Fig. 3B). Twelve hours after infection, gRNA production over the 2-h pulse period was on average 2-fold higher in the SINV-infected untreated cells than in the antibody-treated cells, although the difference was not statistically significant (P = 0.25) (Fig. 3B). By 24 h, the amounts of gRNA were comparable. By 36 h after infection, gRNA production was 1.5-fold higher in the antibody-treated group than in the untreated group (P = 0.11). However, by 48 h, gRNA production declined in both the antibody-treated and untreated cells and the levels were roughly equivalent. sgRNA production over the 2-h pulse period was 2-fold lower at 12 h after infection in the antibody-treated versus untreated cells (P = 0.18), and at 24 h, production was 1.2-fold lower. By 36 and 48 h, levels of synthesis of sgRNA were comparable. To assess the relative proportions of viral RNAs being synthesized, the ratios of sgRNA to gRNA band density at each time point were compared between SINV-infected, untreated, and antibody-treated cells (Fig. 3C). While the ratios were comparable for the early time points, later time points showed that antibody treatment resulted in less sgRNA production. By 48 h after infection, the ratio of sgRNA to gRNA was 2.6-fold lower than in the untreated cells (P < 0.0001). To better quantify changes in RNA synthesis induced by antibody treatment, cells were prelabeled with 5-ethynyl uridine for 15 h. RNA was extracted and nascent RNA isolated at the times indicated by reaction with biotin azide and streptavidin bead capture. Bead-captured viral RNAs were quantified by quantitative reverse transcription-PCR (qRT-PCR) using nsP2 primers (gRNA) and E2 primers (gRNA plus sgRNA), and ratios of nsP2 and E2 RNA expression were compared for treated and untreated infected cells (Fig. 3D). At early times, the ratios were similar, but at 24 h after infection, nascent sgRNA was more abundant in untreated cells than antibody-treated cells. These data recapitulate the trends observed of decreased sgRNA production with antibody treatment.
The nanopore sequencing data identified DVG-like SINV RNAs containing the nsP1 gene that were 1 to 2 kb in length and increased in abundance over time (Fig. 2B). To further characterize these RNAs, a PCR approach was used with primers against the SINV 5′ (SV171F) and 3′ (SV11655R) regions of the genome (Fig. 4). dAP-7 cells were infected with SINV (BHK-21 MOI of 10), and at 4 h after infection were treated with medium (mock) or 5 μg/mL anti-E2 antibody. At 0, 6, 12, 18, 24, 36, and 48 h after infection, total cellular RNA was collected and reverse transcribed to cDNA for PCR using random primers. The RT-PCR yielded two predominant bands: the first was ~11.7 kB (the size of full-length SINV gRNA), and the second was ~1.7 kB (the expected size of the nsP1 DVG) (Fig. 4A). The levels of the 11.7-kb SINV genome were higher and more sustained in the antibody-treated cells, while the levels of the 1.7-kb fragment were overall lower, consistent with trends of increased gRNA and decreased nsP1 DVG with antibody treatment observed in the sequencing (Fig. 2B) and RNA pulse-labeling (Fig. 3) experiments. To confirm the identity of the PCR bands, 4 bands (the 2 predominant bands as well as 2 additional faint bands at 3 kb and 0.9 kb) from the untreated and antibody-treated 6- and 12-h lanes were excised and submitted for Sanger sequencing (Fig. 4A). Three of the bands (11.7, 1.7, and 0.9 kb) yielded sequencing results. The 11.7-kb band matched the full-length SINV genomic RNA. The remaining two bands were nsP1 gene-containing truncated SINV RNAs that matched DVG and subDVG sequences present in the nanopore sequencing data set. The most common and longer of the two DVGs included the nsP1 gene, a portion of the nsP2 gene, the 3′ UTR, and a poly(A) tail (Fig. 4B). The second nsP1 viral RNA was shorter and consisted of most of the nsP1 gene, the 3′ UTR, and a poly(A) tail. The presence of two nsP1 DVGs is consistent with the curved step-down shape of the nsP1 region peak in the sequencing coverage plots (Fig. 2B).
To quantify levels of DVG and full-length SINV genomic RNA during infection, we designed a BRYT Green-based qPCR assay using primer sets that either span the deleted portion of the SINV genome for DVG amplification or are within the deleted portion (nsP3 gene) for SINV genomic RNA detection (Fig. 4C). Using this assay, we evaluated how antibody affects DVG and gRNA levels in BHK-21 cells, undifferentiated cycling AP-7 (cAP-7) cells that model immature neurons, and differentiated AP-7 (dAP-7 cells). While anti-E2 antibody decreases virus production and improves viability in cell types that survive infection with SINV, such as differentiated neuronal cells (e.g., dAP-7 cells) or AT3-bcl-2 adenocarcinoma cells, it is not protective in cell types such as BHK-21, which are interferon-deficient, highly susceptible to SINV infection, and undergo rapid cell death following infection (11, 12, 15, 19). In cAP-7 cells, which express levels of innate immune genes such as IRF3 and IRF7 intermediate between BHK-21 and dAP-7 cells (20), SINV grows to intermediate titers, and antibody has some efficacy, but less than in dAP-7 cells. To determine whether the effect of anti-E2 antibody on nsP1 DVG production is host cell type specific and potentially related to its antiviral effects, BHK-21, cAP-7, and dAP-7 cells were infected with SINV (MOI of 10) and treated with anti-E2 antibody or media (untreated) at 4 h after infection. Twenty-four hours after infection, total cellular RNA was collected, transcribed to cDNA using random primers, and assessed for DVG and genomic RNA via qPCR. Levels of nsP1 DVG were evaluated as fold regulation of nsP1 DVG relative to SINV genomic RNA as determined by the threshold cycle (2−ΔΔCT) method (24). Infections of all three cell types generated detectable levels of nsP1 DVGs that correlated with the viral life cycle kinetics and levels of virus production. BHK-21 cells generated the highest relative amount of nsP1 DVG followed by cAP-7 and finally dAP-7 cells (Fig. 4D). Antibody treatment did not affect nsP1 DVG levels in the BHK-21 cells but significantly decreased nsP1 RNA production in both the cAP-7 and dAP-7 cells (Fig. 4D). These data confirm that nsP1 DVGs are products of the SINV life cycle and further suggest that the effects of anti-E2 antibody on nsP1 DVG production are cell type specific. It is possible, however, with faster growth of SINV in BHK-21 cells that earlier antibody treatment of infected BHK-21 cells may produce more comparable results. Because nsPs are translated from gRNA as a polyprotein, they are present in equal amounts after processing. To determine whether the ratio of nsP1 to the other nsPs is affected by translation of the nsP1 DVG, lysates of infected dAP-7 cells with and without antibody treatment were probed for expression of nsP1, nsP2, and nsP3 (Fig. 4E). The production of all nsPs was suppressed by antibody treatment, but the ratio of nsP1 to nsP2 in untreated cells was greater than those in treated cells at 24 and 48 h after infection (Fig. 4F), suggesting that the nsP1 DVG may be translated.
DVGs are incomplete forms of the viral genome produced during infection due to errors in the viral genome replication process (25). The most common type of DVG involves deletion of an internal portion of the viral genome, as observed with the nsP1 SINV RNA (26, 27). DVGs can be propagated by complementation with wild-type virus and often replicate more rapidly than complete genomes due to their shorter length (28). In addition, if the DVG contains the viral genome packaging signals for association with capsid protein, it can be packaged into defective interfering particles (DIPs) and be transmitted to other cells for further propagation (28). The SINV packaging signal required for viral RNA interaction with the capsid protein for assembly into viral particles includes nucleotides 945 and 1076 of the nsP1 gene (29), which are present within both nsP1 DVG sequences identified. To determine whether the nsP1 DVGs are packaged and possibly transmitted from cell to cell, we examined whether the nsP1 viral RNA is present in extracellular viral particles. dAP-7 cells were infected with SINV and treated with medium only (untreated) or medium containing anti-E2 antibody 4 h after infection, as previously described. Eighteen hours after infection, RNA was extracted from the cellular lysates and cell-free supernatant fluids for RT-PCR detection of the nsP1 DVGs using the same primers as before (SV171F and SV11615R) (Fig. 4A). The nsP1 DVG was detectable in both the cellular and supernatant fluid RNAs from the untreated and antibody-treated cells but was present at a lower level with antibody treatment (Fig. 5A). These data showed that the nsP1 viral RNA is released into the supernatant fluid during infection and confirmed the inhibitory effect of anti-E2 antibody on nsP1 DVG production. To further confirm whether the nsP1 viral RNAs are encapsidated into viral particles, we treated culture supernatant fluids from SINV-infected and antibody-treated cells with RNase A to degrade unprotected single-stranded RNA leaving only encapsidated RNA. Following RNase treatment, RNA was extracted with TRIzol reagent that inhibits RNase activity for analysis. As a positive control for RNase activity, the TRIzol-extracted RNA was also directly treated with RNase A under the same conditions and then reextracted with TRIzol. The final isolated RNA was then assessed for nsP1 DVG using the RT-PCR approach from before. For both the untreated and antibody-treated supernatant fluid samples, RNase A treatment did not completely remove all detectable nsP1 RNA, indicating that some of the nsP1 viral RNA was packaged (Fig. 5B). The positive control (RNase treatment after RNA isolation), however, showed complete nsP1 degradation, confirming that the nsP1 DVG is susceptible to RNase A degradation. Interestingly, RNase treatment did not affect levels of nsP1 RNA in the antibody-treated cells, suggesting that all released nsP1 RNA is encapsidated. This could occur by antibody prolonging survival of infected dAP-7 cells, thereby delaying deterioration of the host cell and release of intracellular viral RNA.
Because the nsP1 DVG sequence can include the SINV 5′ terminus, 3′ UTR, and a poly(A) tail, components necessary for recognition by the host translation machinery, we next asked whether the DVGs could be translated into a viral protein. To answer this question, nsP1 DVG cDNA sequences were cloned into the pcDNA3.1/V5-His-TOPO expression vector with an upstream SP6 promoter for in vitro transcription and protein expression. Only the shorter of the two nsP1 DVGs (subDVG) without a poly(A) tail was successfully cloned due to Escherichia coli toxicity. This DNA sequence was in vitro transcribed to RNA, and a poly(A) tail was subsequently added using poly(A) polymerase. The resulting RNA was transfected into uninfected undifferentiated cAP-7 cells that are more efficiently transfected than dAP-7 cells. RNA transcribed from the pTE plasmid that contains the complete cDNA sequence of SINV TE was transfected in parallel as a positive control, and the nsP1 RNA sequence without a poly(A) tail was included as a negative control. Twenty-four hours after transfection, protein lysates were collected and evaluated by immunoblotting using polyclonal antibody against SINV nsP1. pTE in vitro-transcribed RNA induced robust expression of SINV nsP1 protein at the expected molecular weight of ~60 kDa (Fig. 6A). The lighter band at the slightly lower molecular weight of ~55 kDa was due to antibody nonspecific reactivity and was also present in mock-transfected cells. The nsP1 DVG RNA without a poly(A) tail (dvgRNA−) did not induce protein expression and was identical to the mock-transfected cells. The polyadenylated nsP1 RNA-transfected cells uniquely produced a protein band detectable at a significantly lower molecular weight (15 kDa) than expected (60 kDa), suggesting that the shorter nsP1 DVG results in production of a truncated form of nsP1. Why the protein differs so much in size from the full-length nsP1 protein despite the DVG containing most of the nsP1 sequence is unclear and may be due to the absence of other components from the SINV genome.
We next explored what biological function the nsP1 DVG protein product might have in SINV infection. Due to the correlation between decreased viral replication and decreased nsP1 DVG production, we hypothesized that the nsP1 DVG may be playing a proviral role that is inhibited by anti-E2 antibody. The SINV nsP1 protein has methyltransferase and guanylyltransferase activity, and in cooperation with nsP2, produces the RNA 5′ type 0 7meGpppA cap required for translation (30, 31). However, not all SINV RNAs are capped. Early in infection of BHK-21 cells, released viral particles contain a high proportion of noncapped SINV gRNAs that are noninfectious. Later in infection, the proportion of released virions containing capped SINV gRNAs increases, resulting in mostly infectious virions (32, 33). We hypothesized that nsP1 DVG production and cell-to-cell transmission via DVG-containing particles could be involved in this change by allowing for greater synthesis of nsP1 protein for RNA capping during late infection. Antiviral antibody treatment, by suppressing nsP1 DVG production and neutralizing nsP1 DVG-containing particles, would lead to less nsP1 protein, decreased SINV RNA capping efficiency, and therefore decreased infectious virus production. To assess whether antibody treatment affects the proportion of viral particles that have capped genomes, we performed an exoribonuclease degradation assay utilizing the enzyme XRN-1, which selectively degrades noncapped RNAs. Virus particles from the supernatant fluids of SINV-infected, untreated, and antibody-treated dAP-7 cells 24 h after infection were purified by ultracentrifugation. The RNA was isolated from the virus particles using TRIzol reagent and treated with XRN-1 for 1 h at 37°C to degrade noncapped RNAs. The remaining XRN-resistant RNAs were extracted using phenol-chloroform and reverse transcribed to cDNA for SINV RNA qRT-PCR analysis. The ratio of SINV nsP2 and E2 RNA with or without XRN-1 treatment was compared. Levels of nsP2 and E2 RNA from the SINV-infected, untreated cells were not affected by XRN-1 treatment (ratio of ~1) (Fig. 6B), suggesting that the majority of the packaged SINV genomic RNA is capped at 24 h after infection in dAP-7 cells. In contrast, with anti-E2 antibody treatment, roughly 50% of the virion RNA was susceptible to XRN-1 degradation (64% decrease in nsP2 relative to untreated cells, P < 0.001; 46% decrease in E2, P < 0.001), suggesting that approximately 50% of the packaged SINV RNA was not capped. To confirm the effects of supplemental nsP1 on genome capping and assess the effects of additional nsP1 expression in antibody-treated cells, dAP-7 cells were transiently transfected with lentiviral vector plasmids expressing nsP1 or the nsP1 DVG (Fig. 6C), infected with SINV and then treated or not with anti-E2 antibody. In cells transfected with either DVG or nsP1, the antibody-mediated reduction in capped genomes was reversed to resemble that of untreated cells, further supporting a role for the nsP1 DVG in increasing SINV RNA capping efficiency (Fig. 6D).
In vitro treatment of SINV-infected neuronal cells with bivalent antibody to the E2 glycoprotein inhibits virus replication, restores cellular functions, blocks virus release, and improves neuron viability (11, 12, 14–16, 19). In this study, we used Oxford Nanopore MinION direct RNA sequencing to analyze changes at the RNA level in SINV-infected differentiated AP-7 rat sensory olfactory neurons that were or were not treated with anti-E2 antibody. Antibody treatment decreased overall levels of viral RNA and altered the proportions of different SINV RNA forms. Genomic RNA levels were relatively higher with antibody treatment, while sgRNA levels were lower, consistent with the previously reported antibody-mediated inhibition of structural protein synthesis (16). In addition, antibody suppressed production of two previously unrecognized DVGs encoding the nsP1 capping protein. The nsP1 DVGs could be translated and packaged into extracellular particles. Assessment of the proportion of noncapped RNAs in extracellular virus particles by XRN-1 degradation revealed that 24 h after infection the majority of RNAs from released SINV particles were capped. With antibody treatment, however, only 50% of the virus particle RNA was capped but exogenous expression of nsP1 improved capping efficiency. These data suggest that the DVGs increase the amounts of nsP1 protein produced to improve viral RNA capping and virion infectivity. In addition, we postulate that anti-E2 antibody inhibits SINV replication by decreasing production of nsP1 DVG-containing particles for cell-to-cell transmission and decreasing production of infectious SINV particles containing capped genomic RNA (Fig. 7). The protective effects of anti-E2 antibody are multifactorial and induced by physical binding of antibody to the surface of infected cells and the intracellular signaling cascade triggered (16). During infection SINV induces loss of membrane potential following inhibition of Na+ K+ ATPase activity (15, 34), shutoff of host protein synthesis due in part to activation of RNA-dependent protein kinase (PKR) (18), and apoptotic cell death (35). Although complete mechanisms are not yet established, antiviral antibody partially reverses or inhibits these effects by restoring intracellular cation concentrations and host protein synthesis. Intracellular Na+ and K+ concentrations affect interaction of the viral nsP4 RNA polymerase with the genomic and subgenomic promoters; therefore, restoration of intracellular cation concentrations results in restored host protein synthesis and improved interferon and innate response factor production (14). In addition, antibody binding to viral E2 protein on the surface of infected cells induces activation of the NF-κB signaling and STAT3 cell survival pathways that initiate cellular responses that inhibit amplification of early virus replication complexes (15, 16). In this study, we observed that not all viral RNAs were similarly affected by antibody treatment, and synthesis of gRNA was transiently improved. This is consistent with previously observed delayed processing of the nonstructural polyprotein with antibody treatment that may prolong transcription of the negative-sense genome template and delay the shift to exclusive production of the positive-sense SINV RNAs (15). Translation of viral gRNA requires the same machinery and regulation as translation of cellular mRNAs (36). Because synthesis of the negative-strand genomic RNA template requires continued production of the nonstructural polyprotein (37), maintaining host protein synthesis would also prolong the synthesis of the negative-strand template leading to greater genomic RNA replication. Although of interest, levels of negative-sense gRNA were not specifically evaluated in the nanopore sequencing analysis because it is not polyadenylated or in the pulse-labeling experiments because it is not distinguishable in size from positive-sense gRNA. Future studies that utilize next-generation techniques such as SLAM-seq (thiol-linked alkylation for metabolic sequencing of RNA) to more accurately quantify RNA synthesis during SINV infection with or without antibody treatment will be necessary to confirm these findings. Furthermore, the mechanism for antibody-mediated modulation of viral RNA synthesis is unknown. It is possible that antibody affects the availability of cellular factors that regulate transcription within neurons required for viral RNA synthesis and will require a more comprehensive analysis of the transcriptional and proteomic changes induced by antibody treatment of infected cells. In contrast to gRNA, antibody decreased production of SINV sgRNA early in infection, consistent with the previously described antibody-induced decrease in structural protein synthesis (16). One possible explanation is that restoration of host protein synthesis with antibody treatment may allow production of host antiviral factors that decrease the efficiency of viral transcription, especially from the subgenomic promoter. High-throughput interactome studies have shown that hundreds of mammalian proteins are capable of binding SINV RNA to regulate both transcription and translation (38, 39). Therefore, studies of antibody-induced changes in expression of these RNA-binding proteins, especially those that bind the subgenomic promoter region, would also be a valuable area of investigation. These studies revealed that SINV infection of BHK-21 cells, as well as dAP-7 neuronal cells, resulted in production of previously unrecognized DVGs consisting of the SINV nsP1 gene, sometimes a portion of the nsP2 gene, the 3′ UTR, and a poly(A) tail. Similar deletion-type DVGs have been reported for many families of RNA viruses, including alphaviruses (26, 27, 40, 41). DVGs are products of errors in the viral transcription process and generally inhibitory to wild-type virus replication due to competition for viral/host resources and activation of the host immune response. However, for some viruses, DVGs facilitate infection or persistence (40). In Sendai virus and respiratory syncytial virus infections, high intracellular levels of DVGs protect infected cells from death by altering the outcome of the mitochondrial antiviral signaling-mediated tumor necrosis factor response (42). With hepatitis C virus infection, coexpression of a wild-type virus and DVG results in increased viral replication, suggesting an important biological role for the DVG (43). In our study, the specific accumulation of the nsP1 DVGs suggests that they selectively replicate and may enhance SINV replication in neurons. Observations that further support the possibility that the nsP1 DVGs may play a proviral role in SINV infection include (i) packaging into enveloped particles with the possibility of cell-to-cell spread, (ii) DVG translation into protein, and (iii) association with increased SINV RNA capping. nsP1 has an amphipathic peptide segment and is palmitoylated for binding to the plasma membrane where it induces formation of spherules that organize replication complexes for viral RNA synthesis. nsP1 oligomerizes to form a dodecameric ring at the neck of the spherules to cap RNAs as they are exported (44–46). Genomic RNA needs to be capped for translation of the nonstructural polyprotein, formation of replication complexes, and protection from degradation, so virions containing noncapped genomes are not infectious (33). However, there may be other functions for noncapped viral RNA as analysis of nsP1 mutants with altered capping efficiency has shown that increased capping decreases virus replication in BHK-21 cells by impairing virion production (47). Antiviral antibody treatment suppressed production of the nsP1 DVG and was associated with a higher proportion of noncapped SINV genomes packaged into viral particles than untreated SINV-infected cells. Transfection of infected antibody-treated dAP-7 cells with vectors expressing nsP1 or the nsP1 DVG restored the proportion of capped to noncapped SINV genomes to those present in untreated SINV-infected cells. However, cell type likely substantially contributes to these observations and the relative roles of capped and noncapped RNAs. SINV replication in mature neurons (e.g., dAP-7 cells) is highly restricted compared to replication in BHK-21 cells or immature neurons and virus replication in BHK-21 cells is not decreased by treatment with anti-E2 antibody, so optimal production of infectious particles may be more important in dAP-7 cells than BHK-21 cells (18, 19). In addition, we hypothesize that antibody may neutralize infectivity of defective particles containing the nsP1 DVG, thereby preventing their entry into and propagation in new cells. In addition to its role in RNA synthesis, nsP1 can induce filopodia to facilitate cell-to-cell virus transmission, a property of potential importance for transsynaptic neuronal spread (48–51). nsP1 can interfere with interferon induction and signaling to influence virulence (52–56), and increased capping efficiency is associated with decreased production of inflammatory cytokines, with consequences for neurovirulence (57–59). In vivo, antibody clears infectious virus from SINV-infected brain and spinal cord neurons of mice within 8 days after infection, but viral RNA is eliminated much more slowly and continues to be detectable for months (9–11, 60). Future studies to determine whether the nsP1 DVG is produced in vivo and, if so, whether levels are regulated by antibody within the CNS will be of interest.
AP-7 rat olfactory sensory neuronal cells immortalized with a temperature-sensitive simian virus 40 (SV40) T antigen (gift from Dale Hunter, Tufts University, Boston, MA) (21) were grown under the permissive conditions of 33°C and 7% CO2 in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Atlanta Biologicals), 100 U/mL penicillin, 100 μg/mL streptomycin, and 2 mM glutamine (Gibco). At 25% confluence, cells were differentiated for 7 days by shifting the cultures to 39°C and 5% CO2 and supplementing the medium with 1 μg/mL insulin, 20 μM dopamine, and 100 μM ascorbic acid (Sigma). Cells were routinely tested for mycoplasma using the MycoAlert PLUS mycoplasma detection kit (Lonza).
Stocks of the TE strain of SINV (5) were generated from RNA transcribed from cDNA, transfected into and assayed by plaque formation on BHK-21 cells. Differentiated AP-7 cells were infected at a BHK-21 cell multiplicity of infection (MOI) of 10 in DMEM–1% FBS for 1 h. IgG from the supernatant fluids of the SV127 antibody hybridoma clone (mouse monoclonal IgG3 against the SINV E2 glycoprotein) (12, 61) was purified by adsorption to protein G and resuspended in phosphate-buffered saline (PBS) (Genscript Hybridoma Cell Culture Antibody Production Services). Four hours after infection, cell cultures were washed to remove unbound virus and treated with 5 μg/mL of antibody diluted in DMEM–1%FBS. Following 1 h of incubation, the antibody was diluted to 1.25 μg/mL by the addition of 3 volumes of DMEM–1%FBS and maintained until sample collection.
Cellular RNA samples were collected by direct lysis in TRIzol reagent (Thermo Fisher Scientific), and supernatant fluid RNA samples were collected by adding 1 mL TRIzol to 100 μL culture fluid. RNA was purified according to manufacturer’s instructions and reverse transcribed into cDNA using (i) the Maxima H minus first strand cDNA synthesis kit (Thermo Fisher Scientific) with oligo(dT) primers for RT-PCR or (ii) a high-capacity cDNA reverse transcription kit (Applied Biosystems) with random primers for qRT-PCR. For PCR amplification of SINV RNA for gel electrophoresis, Phusion polymerase (Thermo Fisher Scientific) and the following primers were used for PCR: 171F (5′-GACCATGCTAATGCCAGAGC-3′) and 11655R (5′-GTTATGCAGACGCTGCGTGG-3′). For sequencing, the band of interest was extracted using the QIAquick gel extraction kit (Qiagen) and submitted to the Johns Hopkins Synthesis and Sequencing Facility. To quantify levels of SINV gRNA versus nsP1 DVG, qRT-PCR was performed on the cDNA samples using the GoTaq master mix (Promega) and the following primers on a 7500 Fast real-time PCR system: for detection of the nsP1 defective genome, 1672F (5′-TCGGAGCAGCATTAGTTGAA-3′) and 11615R (5′-ATTATGCACCACGCTTCCTC-3′); for detection of the SINV genomic RNA, 4449F (5′-GGAAAAGACCGCCTTGAAGT-3′) and 4585R (5′-ACAGACTCCTTAAGTTGGAGTGC-3′). To quantify levels of the SINV nsP2 and E2 genes, qRT-PCR was performed using EagleTaq universal master mix (Roche) and the following sequences: for SINV nsP2, nsP2 3373F, 5′-CCG CAA GTA TGG GTA CGA TCA-3′, and nsP2 3454R, 5′-GTG CCC TTC CCA GCT AGC T-3′; for TaqMan probe nsP2 3317, 5′-FAM (6-carboxyfluorescein)-CCA TTG CCG CCG AAC TCT CCC-TAMRA (6-carboxytetramethylrhodamine)-3′; for SINV E2, E2 8732F, 5′-TGG GAC GAA GCG GAC GAT AA-3′, and E2-8805R, 5′-CTG CTC CGC TTT GGT CGT AT-3′; and for TaqMan probe E2 8760, 5′-FAM-CGC ATA CAG ACT TCC GCC CAG T-TAMRA-3′.
For direct sequencing of the RNA, poly(A) RNAs were isolated from frozen cell pellets using TRI-Reagent (Invitrogen), following the manufacturer’s instructions, with 1-bromo-3-chloro-propane for phase separation. Five to 50 μg of total cellular RNA diluted in 50 μL of nuclease-free water was poly(A) enriched using the NEXTflex poly(A) beads (BIOO Scientific). The resulting poly(A) RNAs were eluted in nuclease-free water and quantified using the Qubit 4 fluorometer (Thermo Fisher Scientific). One hundred forty to 800 ng of RNA was prepared for nanopore direct RNA sequencing following the ONT SQK-RNA002 kit protocol, including the optional ONT recommended reverse transcription step using Superscript IV (ThermoFisher). RNA sequencing was performed on the Oxford MinION platform using ONT R9.4.1 flow cells. The ONT Guppy workflow (version 4.0.11) was used for RNA base calling. Strand reads that had an average sequence quality of 7 or higher were classified as passing quality reads. Minimap 2 version 2.17 was used to align the nanopore RNA reads to the rat genome Rnor_6.0 and SINV TE strain genome (NC_001547.1). The junction sites of subgenomic RNA and the DVGs were determined to be positions 7609 and 1751, respectively, and the abundance of each species was defined as the coverage difference at these junctions. The abundance of genomic RNAs for each sample was defined as the coverage at position 1752. Because the subDVG does not have a sharp 3′ junction, the abundance of subDVG RNA is calculated as the difference between the coverage at position 1110 and the minimum coverage between 1111 and the DVG junction.
For radioactive isotope labeling of newly synthesized RNA, dAP-7 cells were incubated at the indicated times after infection with DMEM containing [5,6-3H]uridine (20 μCi/mL) (Perkin Elmer) and dactinomycin (1 μg/mL) (Sigma) at 39°C for 2 h. Cellular lysates were collected with TRIzol and RNA purified as described above. RNA concentrations were determined by NanoDrop spectrophotometer. Two-microgram RNA samples were prepared for electrophoresis with RNA gel loading dye (Thermo Fisher Scientific) and separated on a 1% formaldehyde-agarose gel. Fluorography was performed with 2.5% 2,5-diphenyloxazole as previously described, and the gels were dried and autoradiographed at −80°C (62). Densitometric analysis was performed using ImageJ software. For alkyne-reactive uridine labeling and subsequent capture of newly synthesized RNA, the Click-iT nascent RNA capture kit (Thermo Fisher) was used. In brief, dAP-7 cells were cultured with DMEM containing 5-ethynyl uridine (0.2 mM) for 15 h, and at indicated times after infection with or without antibody treatment, RNA was isolated and biotin azide was clicked onto nascent RNA for capture with streptavidin. One microgram of RNA was heated at 68 to 70°C for 5 min and incubated with Dynabeads MyOne streptavidin T1 for 30 min at room temperature with low vortexing. Beads were captured with a magnet and washed twice before cDNA synthesis with the SuperScript VILO kit. (Invitrogen) and quantitative PCR (qPCR) for gRNA (nsP2) and gRNA plus sgRNA (E2).
Supernatant fluid samples from dAP-7 cells infected with SINV and treated or untreated with anti-E2 antibody were centrifuged at 1,200 × g for 5 min to pellet cellular debris. Cell-free supernatant fluids were collected and treated with 10 μg/mL RNase A (Thermo Fisher Scientific) for 1 h at room temperature. Following treatment, RNA was isolated from the samples with TRIzol reagent and reverse transcribed to cDNA using Maxima H minus first strand cDNA synthesis kit with oligo(dT) primers and assessed by PCR as described above.
The nsP1 defective genome was PCR-amplified from the cDNA of SINV-infected dAP-7 cells 24 h after infection using Easy-A PCR cloning polymerase (Agilent Technologies) and the primers SINV 1F (5′-ATTGACGGCGTAGTACACACTA-3′) and oligo(dT)18. The PCR products were separated by gel electrophoresis, and the band of interest was excised and purified using the QIAquick gel extraction kit (Qiagen). Following sequencing confirmation, the defective genome sequence was cloned into the pcDNA3.1/V5-His-TOPO vector (Thermo Fisher Scientific) and the plasmid was amplified. For production of RNA for transfection, the plasmid was linearized using XhoI (New England Biolabs) and in vitro transcribed and capped using the mMessage mMachine SP6 transcription kit (Ambion). The transcribed RNA was precipitated with LiCl and polyadenylated using poly(A) polymerase (New England Biolabs). The RNA was repurified by phenol-chloroform extraction and ethanol precipitation prior to transfection. For transient transfections, pGenLenti plasmids were generated for transduction to express either nsP1 or DVG (Genscript). dAP7 cells were established as above and incubated with lipofectamine and 1 μg of plasmid for 24 h before confirmation of transcription and translation by qPCR and immunoblotting. Transfected cells were then infected followed by antibody treatment or not as described. Cell lysates were collected in TRIzol for quantification of E2 and nsP2 RNA by qPCR and in radioimmunoprecipitation (RIPA) buffer for protein by immunoblotting.
A total of 106 undifferentiated AP-7 cells were transfected with 2.5 μg RNA and 10 μL of Lipofectamine 2000 reagent (Thermo Fisher Scientific). Twenty-four hours after transfection, protein lysates were collected for immunoblot analysis with radioimmunoprecipitation (RIPA) buffer (50 mM Tris-Cl [pH 8.0], 150 mM NaCl, 1% NP-40, 0.1% SDS, 0.5% Na-deoxycholate, 1 mM EDTA) containing protease and phosphatase inhibitor cocktails (Roche) as previously described (18). Protein concentration was determined using the DC assay kit (Bio-Rad). Ten micrograms of total protein were separated by SDS-PAGE, transferred to a nitrocellulose membrane (Bio-Rad), and blocked in Tris-buffered saline containing 0.1% Tween 20 (TBST) and 5% milk. Immunoblot detection was performed using primary rabbit polyclonal antibodies to nsP1, nsP2, and nsP3 (1:1,000) (63) and secondary horseradish peroxidase-conjugated goat anti-rabbit antibody (1:1,000) (Cell Signaling Technology). Membranes were developed using the Amersham ECL Prime Western blotting detection reagent (GE Healthcare).
Extracellular virus particles were isolated from cell culture supernatant fluids by one-step sucrose cushion ultracentrifugation, as previously described (26), or using the Viraffinity virus and viral component isolation kit (BSG). For sucrose sedimentation, 8 mL of 20% sucrose in PBS was overlaid with 22 mL of the virus-containing medium and centrifuged at 107,000 × g for 2 h at 4°C. For Viraffinity, virus-containing medium was added at a 4:1 ratio to the water-insoluble elastomeric polyelectrolyte and then microcentrifuged. Virus pellets were directly lysed with TRIzol reagent according to the manufacturer’s instructions for RNA analysis. Following purification, 200 μg of RNA was treated with 1 U of 5′- to 3′-exoribonuclease XRN-1 (New England BioLabs) for 1 h at 37°C to degrade noncapped RNAs. The treated RNAs were repurified by TRIzol extraction and reverse transcribed to cDNA using an Applied Biosystems high-capacity cDNA reverse transcription kit with random primers. The cDNA was assessed for SINV nsP2 and E2 gene abundance by qRT-PCR as described above.
Data are expressed as means ± standard deviation (SD). Statistical analyses were performed using Prism Software v8.3 (GraphPad), with a P value of <0.05 being considered significant. Differences between two groups were analyzed by unpaired, two-tailed Student's t test. Multiple comparisons between groups were made using two-way analysis of variance (ANOVA) with Sidak’s posttest.
Data related to this article were deposited in SRA under BioProject no. PRJNA720710. All code can be found at https://github.com/mbartl13/mbartl13 (mbartl13/GriffinLab-Sindbis). | true | true | true |
PMC9600662 | 35993747 | Kristopher J. Kennedy,Florian J. Widner,Olga M. Sokolovskaya,Lina V. Innocent,Rebecca R. Procknow,Kenny C. Mok,Michiko E. Taga | Cobalamin Riboswitches Are Broadly Sensitive to Corrinoid Cofactors to Enable an Efficient Gene Regulatory Strategy | 22-08-2022 | Bacillus subtilis,RNA biology,cobalamin,coenzyme,cofactor,corrinoid,gene regulation,metabolism,physiology,riboswitch,vitamin B12 | ABSTRACT In bacteria, many essential metabolic processes are controlled by riboswitches, gene regulatory RNAs that directly bind and detect metabolites. Highly specific effector binding enables riboswitches to respond to a single biologically relevant metabolite. Cobalamin riboswitches are a potential exception because over a dozen chemically similar but functionally distinct cobalamin variants (corrinoid cofactors) exist in nature. Here, we measured cobalamin riboswitch activity in vivo using a Bacillus subtilis fluorescent reporter system and found, among 38 tested riboswitches, a subset responded to corrinoids promiscuously, while others were semiselective. Analyses of chimeric riboswitches and structural models indicate, unlike other riboswitch classes, cobalamin riboswitches indirectly differentiate among corrinoids by sensing differences in their structural conformation. This regulatory strategy aligns riboswitch-corrinoid specificity with cellular corrinoid requirements in a B. subtilis model. Thus, bacteria can employ broadly sensitive riboswitches to cope with the chemical diversity of essential metabolites. | Cobalamin Riboswitches Are Broadly Sensitive to Corrinoid Cofactors to Enable an Efficient Gene Regulatory Strategy
In bacteria, many essential metabolic processes are controlled by riboswitches, gene regulatory RNAs that directly bind and detect metabolites. Highly specific effector binding enables riboswitches to respond to a single biologically relevant metabolite. Cobalamin riboswitches are a potential exception because over a dozen chemically similar but functionally distinct cobalamin variants (corrinoid cofactors) exist in nature. Here, we measured cobalamin riboswitch activity in vivo using a Bacillus subtilis fluorescent reporter system and found, among 38 tested riboswitches, a subset responded to corrinoids promiscuously, while others were semiselective. Analyses of chimeric riboswitches and structural models indicate, unlike other riboswitch classes, cobalamin riboswitches indirectly differentiate among corrinoids by sensing differences in their structural conformation. This regulatory strategy aligns riboswitch-corrinoid specificity with cellular corrinoid requirements in a B. subtilis model. Thus, bacteria can employ broadly sensitive riboswitches to cope with the chemical diversity of essential metabolites.
Controlling gene expression is an essential task cells accomplish in a variety of ways. Noncoding RNAs are one such means of gene regulation, acting in parallel or in concert with historically better-studied protein-based mechanisms (1). In bacteria and archaea, riboswitches are a widespread type of gene regulatory RNA with the distinct ability to sense particular intracellular metabolites by direct binding (2). These RNAs are typically located in the 5′-untranslated region of mRNA transcripts and function as cis-regulators of downstream genes within their transcripts. A riboswitch is composed of an effector-binding aptamer domain and an expression platform. The aptamer domain adopts a three-dimensional (3D) structure that can bind its cognate effector molecule. The expression platform domain is a regulatory switch that interprets the effector-binding state of the upstream aptamer typically to promote or disrupt the transcription or translation of downstream genes (3). The diversity of riboswitch effectors and regulatory mechanisms has revealed fundamental insights into how bacteria sense and respond to dynamic environments and has also driven new approaches for precise control and manipulation of microbes for human purposes (4–7). Cobalamin (Cbl) riboswitches (also called “B12 riboswitches” or “adenosylcobalamin riboswitches”) are among the most widespread and structurally diverse types of riboswitch in bacteria (8). They directly bind various forms of the enzyme cofactor Cbl as a cognate effector (Fig. 1A and C) (9) to regulate genes involved in the biosynthesis, transport, and usage of Cbl. Cbl-dependent enzymes function in common metabolic pathways, including methionine synthesis, deoxyribonucleotide synthesis, tRNA modification, and the degradation of certain amino acids, fatty acids, and biopolymers (10–23). Cbl is also required for rarer metabolic processes involved in antibiotic synthesis, mercury methylation, catabolism of steroids, and many others (24–37). Comparative genomic studies indicate most bacteria perform Cbl-dependent metabolism and Cbl-riboswitches often regulate these processes (8, 38, 39). However, an overlooked facet among most riboswitch studies is that Cbl is just one member of a class of enzyme cofactors known as corrinoids (Fig. 1B) (40). In fact, Cbl-dependent enzymes in bacteria often function with corrinoids other than Cbl. Yet, it remains unclear whether the dozens of naturally occurring corrinoid cofactors are also Cbl-riboswitch effectors (41–47). Corrinoid cofactors contain a highly substituted corrin ring with a central cobalt ion, a variable “upper ligand” moiety coordinating the β axial face of the cobalt, and a tail structure extending from the corrin ring and terminating in a variable “lower ligand” moiety that often coordinates the α axial face of the cobalt (Fig. 1A and B). The molecular basis of selectivity of Cbl-riboswitches for upper ligand variants of Cbl has been relatively well studied, but selectivity for corrinoid tail variants remains mostly unexplored (9, 48, 49). To our knowledge, only one study has directly examined corrinoid tail specificity of a single Cbl-riboswitch. In vitro binding measurements showed the aptamer of the Escherichia coli btuB Cbl-riboswitch binds the complete corrinoids Cbl and 2-methyladeninylcobamide ([2-MeAde]Cba) with a 3.2-fold difference in affinity (KD = 89 and 290 nM, respectively). Furthermore, cobinamide (Cbi), an incomplete corrinoid with a truncated tail, binds the aptamer with roughly 8,000-fold lower affinity than Cbl (KD = 753 μM), suggesting this aptamer binds corrinoids in a selective manner (50). In light of these previous studies of corrinoid-specific metabolisms and Cbl-riboswitches, we hypothesize Cbl-riboswitches harbor a range of distinct corrinoid tail-specific activities. Here, we examined how a panel of 38 Cbl-riboswitches derived from 12 bacterial species responds to the distinct tail structures of four corrinoids: Cbl, pCbl, and CreCba, representatives of the benzimidazolyl, purinyl, and phenolyl cobamides, respectively, and Cbi, an incomplete corrinoid (Fig. 1B). To compare activities among several dozen Cbl-riboswitches, we devised a live cell fluorescence-based reporter system in Bacillus subtilis. In contrast to conventional in vitro biochemical approaches, this riboswitch reporter system captures the complete corrinoid-responsive gene regulatory process and provides rapid functional measurements with multiple effectors in parallel. Our results obtained from experiments in the reporter system in conjunction with comparative structural analyses of Cbl-riboswitches and corrinoid effectors allowed us to develop a mechanistic model for how corrinoid tail-specific gene regulation is achieved. Additionally, we examine a gene regulatory strategy for the corrinoid specificity of a Cbl-riboswitch and discuss the conceptual and practical implications of these findings.
In order to compare corrinoid selectivity among several Cbl-riboswitches and corrinoids, we constructed an in vivo green fluorescent protein (GFP) reporter system. We initially attempted to use an E. coli host for the reporter system but found most of the riboswitches we tested did not function in the E. coli host. We chose B. subtilis as an alternative host organism because of the robust genome engineering and gene expression toolsets available. Furthermore, the B. subtilis genome does not contain any annotated corrinoid biosynthesis or remodeling genes that would potentially interfere with a riboswitch reporter assay. We engineered the strain to overexpress the Cbl uptake and adenosylation operon, btuFCDR, and deleted queG, which encodes the only Cbl-dependent enzyme in the genome. We found btuFCDR overexpression increased uptake of not only Cbl, but also pCbl, CreCba, and Cbi (Fig. 2A to D). Additionally, each corrinoid was transformed from the cyanated to adenosylated form, suggesting the corrinoids are internalized to the cytoplasm, and not simply accumulating on the outer cell surface. This strain is effective for measuring the response of the B. subtilis btuF Cbl-riboswitch to a broad range of concentrations of Cbl (Fig. 2E to G). Notably, deletion of btuR, which encodes the adensosyltransferase that installs the Ado upper ligand group, rendered the B. subtilis btuF Cbl-riboswitch reporter insensitive to exogenously supplied CNCbl, MeCbl, and OHCbl, while retaining dose-dependent repression in response to AdoCbl (Fig. S1). This strongly suggests the B. subtilis btuF Cbl-riboswitch only responds to Cbl containing the 5′-deoxyadenosine upper ligand, in contrast with a report suggesting this riboswitch aptamer can also bind MeCbl and OHCbl (51). 10.1128/mbio.01121-22.1 The B. subtilis btuF Cbl-riboswitch is sensitive to Cbl with 5’-deoxyadenosyl, but not methyl, hydroxo, or cyano upper ligand moieties. Dose responses of B. subtilis btuF Cbl-riboswitch reporter strains with (A) adenosylcobalamin (AdoCbl); (B) methylcobalamin (MeCbl); (C) hydroxocobalamin (OHCbl); and (D) cyanocobalamin (CNCbl). Data points and error bars represent mean and standard deviation of four independent replicates. Horizontal dotted lines demarcate no change in expression. Note that the two dose-response curves in panel A are overlapping. Download FIG S1, PDF file, 0.03 MB. Copyright © 2022 Kennedy et al. 2022 Kennedy et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
In the strain background described above, we constructed 86 reporter strains to examine riboswitches from 20 bacterial species, including 10 species known to produce or require specific corrinoids. Of the 86 reporters, 38 repressed GFP expression 0.5-fold or greater in response to one or more corrinoids. Thirty-seven of these 38 functional riboswitch expression platforms contain a predicted intrinsic transcriptional terminator suggesting they are transcriptional riboswitches. We observed extensive variation in sequence length and nucleotide composition throughout the aptamers and expression platforms of the 38 riboswitches that were functional in B. subtilis (Fig. 1D). Nine of the functional riboswitches are from Priestia (formerly Bacillus) megaterium, which produces Cbl (41), and 12 are from Sporomusa ovata and Veillonella parvula which both produce CreCba (44, 45, 52, 53). These results show Cbl-riboswitches of diverse sequence composition and origin can be examined with the in vivo reporter system. To address whether Cbl-riboswitches are corrinoid selective, and how corrinoid selectivity varies among Cbl-riboswitches, we measured the dose responses of the 38 functional Cbl-riboswitches to four corrinoids, Cbl, pCbl, CreCba, and Cbi. Our results show all of the riboswitches responded to more than one corrinoid, and a subset responded to all four (Fig. 3). Strikingly, all of the tested riboswitches are either semiselective (responding to more than one corrinoid) or promiscuous (responding to all four corrinoids) (Fig. 3A and B). We did not find any highly selective riboswitches that respond to only one corrinoid. The semiselective and promiscuous riboswitches all respond to Cbl and pCbl (Fig. 3C), and the promiscuous riboswitches additionally respond to Cbi and CreCba (Fig. 3D and E, points above the horizontal dashed line). Furthermore, the semiselective riboswitches are generally more sensitive to Cbl than to pCbl, while the promiscuous riboswitches respond similarly to these two corrinoids. Almost all of the riboswitches respond weakly to CreCba compared with the other three corrinoids (Fig. S2A to C). In general, corrinoid selectivity of a riboswitch appears to be associated with its taxonomic origin (Fig. 3C to E). The riboswitches from the Bacilli class are exclusively semiselective (Fig. S2A), whereas those from Negativicutes are predominantly promiscuous (Fig. S2B). The S. ovata cobT riboswitch is a notable exception discussed later. In contrast to taxonomy, corrinoid selectivity of a riboswitch is not strongly associated with the function of its regulatory target genes (Fig. S3). 10.1128/mbio.01121-22.2 Corrinoid dose responses of riboswitch reporter strains. Riboswitch names (species and downstream gene) are indicated for each dose response. Data points are plotted for at least two independent experiments for each strain. Lines connect mean values. (A) Riboswitches from the class Bacilli. Species include Bacillus subtilis, Priestia megaterium, and Alkalihalobacillus halodurans. (B) Riboswitches from the class Negativicutes. Species include Sporomusa ovata, Veillonella parvula, and Selenomonas sputigena. (C) Riboswitches from the classes Clostridia, Deltaproteobacteria, and Bacteroidia. Clostridioides difficile, Clostridium novyi, Desulfitobacterium hafniense, and Symbiobacterium thermophilum are members of Clostridia. Desulfobulbus propionicus and Pelobacter propionicus are members of Deltaproteobacteria. Bacteroides thetaiotaomicron is a member of Bacteroidia. Download FIG S2, PDF file, 0.1 MB. Copyright © 2022 Kennedy et al. 2022 Kennedy et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. 10.1128/mbio.01121-22.3 Cbl-riboswitch specificity does not cluster by regulatory gene target function. Pairwise comparisons of GFP fold repression induced by 100 nM doses of Cbl versus Cbi (top left) pCbl (top right), and CreCba (bottom left). Data points are colored by predicted function of downstream regulatory target genes. The number of riboswitches analyzed from each group is indicated in parentheses. Vertical and horizontal gray dashed lines demarcate a lack of response to one corrinoid. Diagonal line indicates equal repression between corrinoids. The distance of a point from the diagonal line indicates the bias in response towards one of the two corrinoids. Download FIG S3, PDF file, 0.2 MB. Copyright © 2022 Kennedy et al. 2022 Kennedy et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. Next, we attempted to identify the RNA sequence features that underlie corrinoid selectivity. Chimeric fusions of the semiselective P. megaterium metE riboswitch and the promiscuous V. parvula mutA riboswitch enabled us to examine the effects of specific domain and subdomain sequences on corrinoid selectivity (Fig. S4). Fusing the P. megaterium metE aptamer domain to the expression platform of the V. parvula mutA riboswitch produced a semiselective riboswitch chimera, while the reciprocal chimera was promiscuous, suggesting the aptamer domain is a major determinant of corrinoid selectivity (Fig. S4A). However, results from aptamer subdomain swaps of stem P1, stem-loop P2-L2, stem-loop P4-L4, and P6 accessory region were less conclusive. In the context of the P. megaterium metE riboswitch scaffold, swapping stem-loop P2-L2 or the P6 accessory region with the corresponding structures of the V. parvula mutA riboswitch produced chimeras that are less selective by gaining sensitivity to Cbi and CreCba (Fig. S4B). This suggests that these subdomains confer corrinoid promiscuity. Yet, within the V. parvula mutA riboswitch scaffold, swapping stem P1 increased corrinoid selectivity by retaining sensitivity to Cbl and pCbl, but losing sensitivity to CreCba and Cbi (Fig. S4C). The remaining chimeras partially or completely lost overall activity. These results demonstrate subdomains distributed throughout the aptamer domain may impact corrinoid selectivity; no single conserved substructure completely controls corrinoid selectivity, nor did any single structure fully convert a riboswitch’s corrinoid selectivity. Thus, the source of the corrinoid selectivity phenotype appears to be complex and requires inputs from multiple subdomains of the aptamer. 10.1128/mbio.01121-22.4 Chimeric riboswitches show multiple components contribute to corrinoid specificity. Repression of GFP expression with 100 nM corrinoid is shown for (A) P. megaterium metE and V. parvula mutA riboswitches and chimeric fusions of their aptamer and expression platform domains. Aptamer subdomains swaps within the (B) P. megaterium metE riboswitch scaffold and (C) V. parvula mutA riboswitch scaffold. Columns and error bars represent mean and standard deviation of 3 replicates. Cartoon riboswitches depict P. megaterium riboswitch sequence in cyan and V. parvula riboswitch sequence in orange. Sequences in black are identical between riboswitches. Download FIG S4, PDF file, 0.6 MB. Copyright © 2022 Kennedy et al. 2022 Kennedy et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
We next sought to identify how structural differences in the corrinoid tail affect the response of semiselective Cbl-riboswitches to corrinoids. There are no predicted hydrogen bond interactions between the lower ligand and the RNA in the X-ray crystal structures of Cbl-bound riboswitches, making it difficult to surmise how a Cbl-riboswitch might distinguish between corrinoids (51, 54, 55). Could the overall structural conformation of the corrinoid, rather than specific interactions between the RNA and the corrinoid tail, influence Cbl-riboswitch activity? Corrinoids undergo major conformational changes when spontaneously switching between two distinct states known as “base-on” and “base-off” (56, 57). In the base-on state, a nitrogen atom in the lower ligand base is coordinated to the central cobalt atom of the corrin ring (as shown in Fig. 1A). In the base-off state, the lower ligand base is decoordinated, allowing the tail to move more freely (58, 59). Benzimidazolyl and purinyl cobamides (e.g., Cbl, pCbl) can switch between base-on and base-off states. However, the tail moieties of phenolyl cobamides (e.g., CreCba) and Cbi cannot coordinate cobalt; therefore, these corrinoids exist exclusively in a decoordinated state (52). Interestingly, we noticed the semiselective riboswitches respond strongly to Cbl and pCbl, but weakly to Cbi and CreCba. We also observe semiselective Cbl-riboswitches are most sensitive to Cbl, which forms the base-on state more readily than pCbl (Fig. 3C; Fig. S2) (58, 59). In line with these results, the E. coli btuB riboswitch aptamer was previously shown to bind Cbl with higher affinity than Cbi and [2-MeAde]Cba (a primarily base-off corrinoid) (50). Also, all six X-ray crystal structures of Cbl-riboswitches contain Cbl in the base-on state (51, 54, 55). Based on these observations, we hypothesized that semiselective riboswitches distinguish between base-on and base-off states of corrinoids. We therefore used a range of corrinoids with diverse lower ligand structures to test whether the activity of Cbl-riboswitches quantitatively correlates with the base-on tendency of corrinoids. We selected a panel of 16 corrinoids for this analysis, including both natural and synthetic benzimidazolyl, purinyl, and azabenzimidazolyl corrinoids. These corrinoids span a range of base-on tendency between that of Cbl and pCbl, which we measured as the ratio of spectral absorbance at 525 and 458 nm in the adenosylated form (Fig. 4A; Fig. S5). Base-on/base-off equilibrium constants for AdoCbl, Ado[2-MeAde]Cba, and pCbl in aqueous conditions have been reported as 76, 0.48, and 0.30, respectively, and are consistent with our measurements of corrinoid base-on tendency (58, 59). We observed a strong association between base-on tendency and riboswitch response in semiselective riboswitches (Fig. 4B). Even among promiscuous Cbl-riboswitches, we observe measurable sensitivity to base-on tendency, albeit to a much smaller degree (Fig. 4C). These results support the hypothesis that Cbl-riboswitches selectively respond to corrinoids by distinguishing between the base-on and base-off states of corrinoids. 10.1128/mbio.01121-22.5 Corrinoid lower ligand structure impacts base-on tendency. Absorbance spectra of corrinoids were measured in a neutral buffered solution (pH 7.3). Absorbance peaks at 458 nm and 525 nm are associated with base-off and base-on conformations, respectively. Base-on tendency was measured as the ratio between absorbance at 525 nm and 458 nm (thin vertical lines). Each panel is labeled with corrinoid name, base-on tendency value (Abs525nm/Abs458nm), and lower ligand chemical structure. Note the tail structures of PheCba, CreCba, and Cbi cannot coordinate cobalt, and thus these corrinoids cannot be assigned a “base-on tendency” per se. The Abs525nm/Abs468nm ratios of PheCba, CreCba, and Cbi simply reflect an absence of any Co-N coordination at the α axial face. For reference, the dashed line in the upper left panel is the absorbance spectrum of nearly complete base-off Cbl measured in acidic solution (pH 1.57). Download FIG S5, PDF file, 0.3 MB. Copyright © 2022 Kennedy et al. 2022 Kennedy et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
The results presented above led us to speculate about how a Cbl-riboswitch could detect the base-on and base-off state of a corrinoid. In all published X-ray crystal structures of Cbl-riboswitches, aptamer-effector binding is achieved mainly through van der Waals forces and shape complementarity between the binding site and base-on Cbl. Only a few hydrogen bonds between the RNA and corrinoid are observed, none of which occur with the lower ligand group of Cbl (51, 54, 55). Thus, it appears unlikely the riboswitch is directly detecting the specific chemical differences among corrinoid lower ligands. Instead, we considered whether the riboswitch discriminates base-on and base-off forms of a corrinoid by sensing corrinoid conformation. In the base-on state, the tail is spatially constrained due to the Co-N coordinate bond, whereas in the base-off form it is able to sample a wider range of spatial positions (60). To develop mechanistic insight into how the base-on and base-off states a corrinoid could impact Cbl-riboswitch activity, we leveraged the plethora of publicly accessible X-ray crystal structures of macromolecule-bound Cbl (37, 51, 54, 55, 61–70). We first assessed the range of structural conformations that are potentially sampled by corrinoids as they dynamically switch between base-on and base-off states by aligning and visually comparing various structural models of Cbl. Six base-on Cbl models were obtained from structural studies of Cbl-riboswitches and synthetic Cbl RNA aptamers, whereas base-off/His-on Cbl models were obtained from X-ray crystal structures of 10 Cbl-dependent enzymes (Table S2). After aligning and superimposing these molecular models by their central cobalt and coordinating nitrogen atoms, we observed the corrin rings and their amide and methyl substituents occupy similar spatial positions, but the tails of base-on and base-off Cbl structures occupy distinct positions (Fig. 5A and B). Moreover, the base-on Cbl tails appear in very similar positions with lower ligands in close proximity to the central cobalt ion, whereas the base-off Cbl tails appear more scattered with lower ligands more distal to the cobalt ion. These structural alignments visually convey the degree to which the conformations of base-on and base-off corrinoids can vary among biomolecular complexes. 10.1128/mbio.01121-22.9 X-ray crystal structural models used for 3D Cbl structural alignments. Download Table S2, DOCX file, 0.01 MB. Copyright © 2022 Kennedy et al. 2022 Kennedy et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. Next, we compared the positions of the aligned base-on and base-off Cbl models in the context of 3D Cbl-riboswitch models. We analyzed X-ray crystal structures of the two Cbl-riboswitches that contain resolved kissing loop structures: one from Thermoanaerobacter tengcongensis (Fig. 5C and D) and one identified from a marine metagenome sequence (Fig. S6A and B) (54). These structural models show the base-on tails are contained within the binding site, whereas the tails of the base-off Cbl structures protrude away from the binding site and clash with the L5-L13 kissing loop. Although the B. subtilis btuF (Fig. S6C and D) and Symbiobacterium thermophilum cblT (Fig. S6E and F) riboswitch models do not contain the L13 structure of the kissing loop, some of the modeled base-off tails clash with L5 of the aptamer in these structures (51, 55). The kissing loop has been shown to play a key mechanistic role of sensing the corrinoid-binding state of the aptamer domain to influence downstream regulatory structures in the expression platform (71, 72). If kissing loop formation in semiselective Cbl-riboswitches is sensitive to the base-on and base-off states of the corrinoid, then corrinoid selectivity may be mediated by either selective binding by the aptamer or selective formation of downstream regulatory structures. 10.1128/mbio.01121-22.6 Cbl-binding sites in X-ray crystal structures of various Cbl-riboswitches. The marine metagenome derived riboswitch env8 (PDB ID 4FRN) (A, B), the B. subtilis btuF Cbl-riboswitch (PDB ID 6VMY) (C, D), and the S. thermophilum cblT Cbl-riboswitch (PDB ID 4GXY) are depicted with base-on (A, C, E) and base-off (B, D, F) Cbl alignments. Cbl models were aligned by the cobalt and coordinating nitrogen atoms in the corrin ring. Structures of the corrin ring, cobalt, and tail of Cbl are colored in black, blue and green, respectively. Upper ligand structures of Cbl were omitted for clarity. Riboswitch RNA structures are depicted as space-filled models with the L5-L13 kissing loop in pink in panels A and B, and just the L5 in pink in panels C to F. The rest of the RNA structures in each panel are in gray. Arrows in panels D and F point to clashes between RNA and the Cbl tails. Download FIG S6, PDF file, 2.1 MB. Copyright © 2022 Kennedy et al. 2022 Kennedy et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. To determine whether the expression platform structures can impact corrinoid selectivity, we examined corrinoid-selective binding separately from subsequent corrinoid-selective regulation. We tested for promiscuous binding by comparing the Cbl dose response of the P. megaterium metE Cbl-riboswitch in the presence and absence of competing 100 nM Cbi and found the response to Cbl is unaffected by Cbi (Fig. S7A and B). This indicates Cbi does not compete with Cbl for riboswitch binding, supporting corrinoid-selective binding as the mechanism of semiselectivity. However, when the aptamer of this semiselective riboswitch is replaced with the aptamer of the promiscuous S. ovata nikA riboswitch, it retains semiselectivity, suggesting the P. megaterium expression platform also plays a role in corrinoid selectivity (Fig. S7A). Interestingly, the Cbl dose response of the S. ovata nikA/P. megaterium metE chimeric riboswitch does become sensitized to competing Cbi addition, confirming the S. ovata nikA aptamer retains sensitivity to base-off corrinoid in the context of this chimera (Fig. S7C). Taken together, these results show base-off corrinoids may impede both Cbl-riboswitch binding and formation of regulatory structures, explaining the link between corrinoid base-on tendency and riboswitch activity observed in Fig. 4. 10.1128/mbio.01121-22.7 The aptamer and expression platform of the P. megaterium riboswitch contribute to corrinoid selectivity. (A) Repression of GFP expression with 100 nM corrinoid is shown for the P. megaterium metE riboswitch, S. ovata nikA riboswitch, and a chimeric fusion of the S. ovata nikA aptamer and P. megaterium metE expression platform domains. Cbl dose responses with or without 100 nM Cbi in (B) P. megaterium metE and (C) chimeric riboswitches. Data points and error bars represent mean and standard deviation of four independent replicates. Download FIG S7, PDF file, 0.2 MB. Copyright © 2022 Kennedy et al. 2022 Kennedy et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
While the prior experiments clearly demonstrate Cbl-riboswitches are capable of distinguishing between corrinoids, we wondered what purpose Cbl-riboswitch corrinoid selectivity might serve in the organisms containing these regulatory systems. We posit Cbl-riboswitch selectivity reflects a regulatory strategy that complements the corrinoid-specific requirements of the cell and avoids gene mis-regulation. As a specific example, we hypothesize only corrinoids that are functionally compatible with a Cbl-dependent enzyme should cause riboswitch-mediated repression of the expression of its Cbl-independent counterpart (Fig. 6A and B). We tested this hypothesis directly in B. subtilis by examining the function and regulation of methionine synthase isozymes MetE (Cbl-independent) and MetH (Cbl-dependent). In bacterial genomes with both metE and metH, an S-adenosyl methionine (SAM) riboswitch fused in tandem with a Cbl-riboswitch is commonly found upstream of the metE gene (73, 74). The B. subtilis genome contains metE but lacks metH, and no Cbl-riboswitch is located upstream of metE. We, therefore, constructed strains of B. subtilis that heterologously express the metE or metH locus from the Cbl-producing species P. megaterium, in a ΔmetE background with overexpressed corrinoid uptake genes (Fig. 6C) (75–77). In each strain, the P. megaterium genes are constitutively transcribed from the promoter PVeg and also contain a transcriptionally fused gfp to measure expression levels. Growth of the B. subtilis strain expressing P. megaterium metH in a medium lacking methionine was supported to various extents by most benzimidazolyl and both phenolyl cobamides, but not by [5-OHBza]Cba, the purinyl cobamides, or Cbi (Fig. 6D, red squares). This result indicates MetH-dependent growth is influenced by the corrinoid tail structure, as observed previously in other bacteria (78–82). In the B. subtilis strain containing the P. megaterium metE locus which includes the repressing SAM-Cbl-riboswitch, growth was suppressed by benzimidazolyl cobamides to different extents (Fig. 6D, blue circles). This growth pattern coincides with the GFP repression measured for the metE riboswitch (Fig. 6E), except that the repression of metE by purinyl cobamides is apparently insufficient to suppress growth in this context. Comparison of the two strains in response to a suite of corrinoids reveals a striking correspondence between riboswitch-mediated suppression of growth in the metE-containing strain and growth promotion by corrinoids in the metH-containing strain (Fig. 6C). [5-OHBza]Cba and the phenolyl cobamides are exceptions to the trend, though in neither case is MetE-dependent growth completely suppressed by a corrinoid incompatible with MetH. This result demonstrates riboswitch-based repression and cobalamin-dependent isozyme function are largely aligned for the P. megaterium metE-metH pair and suggests that Cbl-riboswitch specificity may generally adhere to a regulatory strategy reflecting the cell’s corrinoid preference.
Riboswitches are key regulators of microbial gene expression. The Cbl-riboswitch was the first type discovered and is among the most widely distributed riboswitch classes in bacteria and archaea (9, 83). Previous biochemical and structural studies have uncovered the major molecular features of the Cbl-riboswitch response to Cbl, including how upper ligand variants of Cbl impact their function (48, 49, 51, 54, 55, 71, 72, 74, 84). Yet few studies have examined how other naturally occurring corrinoids containing diverse lower ligand structures impact gene regulation by Cbl-riboswitches (50). Here, we found Cbl-riboswitches vary in their ability to discriminate between corrinoids, with some being semiselective on the basis of corrinoid base-on/off state, and others being promiscuous. These results were enabled by a carefully designed fluorescent reporter system capable of measuring the responses of dozens of Cbl-riboswitches to multiple corrinoids in vivo (Fig. 2; Fig. S1). Because several naturally occurring corrinoids other than Cbl appear to be potent effectors for Cbl-riboswitches, we propose the term “corrinoid riboswitch” be adopted to describe this broad class of RNAs more accurately. This may also mitigate the inconsistent and overlapping terminology used in the literature (i.e., cobalamin riboswitch, adenosylcobalamin riboswitch, B12 riboswitch, vitamin B12 riboswitch, etc.). We can roughly estimate the intracellular corrinoid concentrations in our corrinoid dose-response experiments to assess the physiological relevance the in vivo Cbl-riboswitch data. Observing most of the corrinoid in the medium is imported by the B. subtilis riboswitch reporter cells (Fig. 2A to D), and assuming a cell volume of 10−15 L and a culture cell titer of 1012 cells/L at OD600 = 1, we calculate the cytoplasmic corrinoid concentrations to range from 0.01 to 100 μM in the corrinoid dose-response experiments—a 1,000-fold increase from the cell culture medium to the cytoplasm. Reports of binding affinity (KD) for riboswitch aptamers to AdoCbl range from 0.026 to 90 μM, which is within the range of our estimated intracellular corrinoid concentrations (48, 50, 54, 72, 85). Additionally, Cbl uptake has been measured in some bacterial species. In E. coli, the minimum cytoplasmic Cbl concentration to support MetH-dependent growth is roughly 0.03 μM (86). In studies of Cbl uptake across several bacterial species, saturating Cbl uptake can result in cytoplasmic Cbl concentrations in the low μM to low mM range, depending on the species (87, 88). Taken together, these data and calculations support the physiological relevance of the Cbl-riboswitch responses measured in this study. We observed Cbl-riboswitches display different degrees of corrinoid selectivity, with some that responded to a subset of corrinoids (semiselective) while others responded to all tested corrinoids (promiscuous) (Fig. 3; Fig. S2). Our chimeric riboswitch results suggest sequence and structural determinants of corrinoid selectivity are dispersed throughout the Cbl-riboswitch aptamer scaffold rather than being confined to a single conserved region (Fig. S4). This finding contrasts with other studies of riboswitch specificity. For example, in a study of Cbl upper ligand specificity, a few key residues in the Cbl binding site were sufficient to fully convert a Cbl-riboswitch from MeCbl-specific to AdoCbl-specific (48). In purine riboswitches, effector specificity is achieved by positioning of a critical conserved uracil or cytosine residue in the binding site of the aptamer, which forms a base-pair with the adenine or guanine effector, respectively. Among the various families of SAM riboswitches, highly specific binding of SAM and exclusion of S-adenosyl homocysteine (SAH) is achieved by RNA structures that discriminate the charged sulfonium ion of SAM from the uncharged sulfoether of SAH (89). In contrast, the SAM/SAH riboswitch class attains effector promiscuity for SAM and SAH by a general lack of interaction between the RNA and the aminocarboxypropyl side chains of these effectors (90, 91). This is reminiscent of Cbl-riboswitches which similarly have few molecular contacts between the RNA and corrinoid tail (50, 54, 55, 84). Overall, our structural model analyses support a mechanism in which semiselective Cbl-riboswitches primarily sense the distinct corrinoid tail orientations of the base-on and base-off forms. In this case, Cbl-riboswitches only indirectly sense the chemical composition of the variable lower ligand group, with differential binding largely determined by steric effects and shape complementarity (Fig. 4 and 5). Recent molecular dynamics simulations of the T. tengcongensis Cbl-riboswitch suggest that the kissing loop structure may form prior to effector binding, which would place even greater constraints on the corrinoid tail orientation to achieve shape complementarity with its binding site (92). Additionally, some Cbl-riboswitches may in fact bind base-off corrinoids but disrupt subsequent formation of downstream regulatory structures of the expression platform, perhaps by interfering with the kissing loop (Fig. S7). A similar feature has been observed in tetrahydrofolate (THF) riboswitches where chemical variations in the para-aminobenzoic acid moiety of THF analogs differentially perturb expression platform structures without affecting aptamer binding (93). The mechanisms of corrinoid selectivity of Cbl-riboswitches could be directly tested in future structural or biochemical studies of promiscuous Cbl-riboswitches with base-on and base-off corrinoids. In regard to gene regulatory strategies, it seems sensible that Cbl-riboswitches are not highly effector-specific because bacteria are often flexible in their corrinoid usage. A variety of corrinoids have been shown to support growth of C. difficile, S. ovata, and Ensifer meliloti despite each of these organisms displaying highly specific corrinoid production (82, 94, 95). Furthermore, because corrinoid auxotrophy is prevalent among corrinoid-dependent bacteria, many organisms may need to take advantage of the wide range of corrinoids that may be available in their environment (96). Thus, the range of effector selectivity we observe among Cbl-riboswitches may reflect a coevolution between corrinoid-responsive gene regulation and corrinoid-dependent physiology. Our result demonstrating complementary corrinoid selectivity between P. megaterium MetH-dependent growth and Cbl-riboswitch-dependent expression of MetE is consistent with this notion (Fig. 6). Alternatively, the preference for base-on corrinoids among Cbl-riboswitches may function as a proxy to discriminate complete corrinoid coenzymes from incomplete corrinoids such as Cbi, which often function poorly as coenzymes. This idea has been proposed as an explanation for the remarkably high selectivity of the corrinoid uptake system in mammals (97). Interestingly, we found all but one of the S. ovata riboswitches tested are promiscuous types that can respond to its natively produced CreCba. The notable exception is a semiselective riboswitch upstream of the gene cobT (Fig. S2B), which functions in a late step of corrinoid biosynthesis that occurs after synthesis of Cbi (53, 98–103). Thus, this Cbl-riboswitch that discriminates against Cbi may allow homeostatic regulation of cobT in response to complete corrinoids like Cbl, while preventing unproductive repression of cobT in the presence of incomplete corrinoids like Cbi. Future studies examining corrinoid-specific gene regulation of riboswitches in the context of their native organisms may help clarify which regulatory strategies are generally at play in corrinoid-related bacterial physiology. Our findings fit into a broader discussion of how corrinoids impact complex microbial communities (104). In future studies, it will be worth examining how the interplay between corrinoid-specific gene regulation and corrinoid-specific metabolic pathways influence microbial interactions. There is also significant interest in the fields of bioengineering and synthetic biology to use riboswitches as gene regulatory devices because they act more rapidly and efficiently than protein-based regulatory systems (6, 7, 105, 106). Riboswitches have also become desirable therapeutic drug targets because they often control essential metabolic pathways in pathogenic microbes (4, 5, 107). Broader consideration of the conformational dynamics of larger types of effector molecules, including organic cofactors, antibiotics, and their analogs could aid efforts to engineer synthetic RNA-based regulatory platforms that function robustly in vivo. This could also inform future efforts to develop synthetic antimetabolites that, for example, elicit gene mis-regulation, instead of a more common strategy of creating inhibitors for enzymes (108, 109). The chemical diversity of corrinoids is intrinsically linked to a vast array of metabolic processes and microbial interactions. Yet it remains unclear how microbes have evolved to cope with and thrive on the assortment of natural corrinoid analogs, especially compared with other primary metabolites, including organic cofactors, nucleotides, and amino acids which typically require one specific structural form for precise biological functions. We have gained new appreciation for the impacts of chemical diversity on biological function by focusing on the Cbl-riboswitch with its distinctively complex structure and regulatory mechanism, and by accounting for the often overlooked biological and ecological roles of corrinoid analogs. Future studies into the evolution of microbial molecular specificity for corrinoids may yield further insight into the nature of these exceptionally versatile coenzymes.
Cbl-riboswitch sequences, chromosomal coordinates, and regulon information were downloaded from the RiboD online database (73) (Table S1). The genome of Sporomusa ovata was not included in the RiboD database, so we used the RiboswitchScanner webserver to search for Cbl-riboswitches in this organism (110, 111). Cbl-riboswitch aptamer sequences were manually aligned by conserved secondary structures bounded by the 5′ and 3′ ends of the P1 stem (8, 85) (Data Set S1). The P13-L13 stem-loop and potential intrinsic transcriptional termination hairpin structures of the expression platform were identified using secondary structure prediction tools in RNAstructure 6.2 (112, 113). Intrinsic terminators were identified as stem loops directly preceding a sequence of five or more consecutive uracil residues (114, 115). Sequence alignment and annotation was carried out in JalView 2.11.1.4 (116). Cartoons of riboswitch secondary structures were constructed using the StructureEditor program of RNAstructure 6.2 (113). 10.1128/mbio.01121-22.8 Strain list and riboswitch informationTable S1, XLSX file, 0.03 MB. Copyright © 2022 Kennedy et al. 2022 Kennedy et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license. 10.1128/mbio.01121-22.10 Riboswitch sequence alignment in Stockholm sequence format. Download Data Set S1, TXT file, 0.03 MB. Copyright © 2022 Kennedy et al. 2022 Kennedy et al. https://creativecommons.org/licenses/by/4.0/ This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
Cyanocobalamin, adenosylcobalamin, methylcobalamin, hydroxocobalamin, and dicyanocobinamide were purchased from MilliporeSigma. All other corrinoids used in this study were produced in bacterial cultures and purified in cyanated form as previously described (82, 95, 117, 118). For the experiments in Fig. 4 and Fig. S5, corrinoids other than cobalamin were chemically adenosylated to obtain the coenzyme (5′-deoxyadenosylated) form as previously described (82, 117). UV/Vis spectra were collected from corrinoid samples in UV/Vis-transparent 96-well microtiter plates (greiner bio-one UV-STAR 675801) using a BioTek Synergy 2 or Tecan Infinite M1000 Pro plate reader. To measure concentrations of corrinoid stock solutions, corrinoid samples were diluted 10-fold in 10 mM sodium cyanide to obtain the dicyanated base-off form of the corrinoid. The concentration of the dicyanated corrinoid was calculated using the extinction coefficient ε580 = 10.1 mM−1 cm−1 (52, 119). For adenosylated corrinoids used in Fig. 4 and Fig. S5, base-on tendency at neutral pH was measured as the ratio of spectral absorbance at 525 nm and 458 nm in phosphate-buffered saline solution pH 7.3 at 37°C (47, 59).
Plasmids generated in this study were constructed with one-step isothermal assembly (120) and introduced into E. coli strain XL1-Blue by heat shock transformation. Riboswitch reporter plasmids were constructed in the shuttle vector pSG29 for single copy integration at the amyE locus of the B. subtilis chromosome (121). Riboswitch DNA sequences were inserted between the transcriptional start site of the constitutive PVeg promoter and the gfp translational start site of pSG29. For riboswitch sequences that resulted in no detectable GFP signal under any conditions, a synthetic ribosome binding site (RBS) sequence R0 (5′-GATTAACTAATAAGGAGGACAAAC-3′) from pSG29 was placed between the riboswitch sequence and gfp translational start site (Table S1B and C). Strains used in this study are listed in Table S1A. All B. subtilis riboswitch fluorescent reporter strains and B. subtilis strains expressing P. megaterium metE and metH used in this study are derived from the high-efficiency transformation strain SCK6, which has a xylose-inducible competence gene cassette (122). Preparation of competent cells and transformations of all SCK6-derived strains were performed as previously described (122). The strain KK642, which constitutively overexpresses the corrinoid uptake genes, was constructed by deletion of gene queG and replacement of the promoter and 5′ untranslated region of the btuFCDR operon with the PVeg promoter and R0 RBS (121). B. subtilis genes queG, btuR, and metE were targeted for deletion by recombination with kanamycin resistance cassettes containing flanking sequence homology to each respective locus. Kanamycin resistance cassettes were PCR-amplified from genomic DNA of B. subtilis strains BKK08910 (ΔqueG::kanR), BKK33150 (ΔbtuR::kanR), and BKK13180 (ΔmetE::kanR) (123). Kanamycin resistance cassettes were removed by Cre-Lox recombination using plasmid pDR244 as previously described (123). B. subtilis strains heterologously expressing metE and metH from P. megaterium were constructed as follows. The metE and metH genes were PCR-amplified from genomic DNA of P. megaterium DSM319 and cloned between the transcriptional start site and the gfp translational start site of pSG29. The P. megaterium metE amplified fragment starts at the SAM-Cbl tandem riboswitch in the 5′ UTR and ends at the metE stop codon, whereas the P. megaterium metH fragment starts at the metH RBS, which is composed of 20 nucleotides preceding the metH translational start site and ends at the metH stop codon. Riboswitch reporter plasmids and plasmids containing P. megaterium metE and metH were linearized by restriction enzyme digest with ScaI-HF (New England Biolabs) and selected for integration at the amyE locus of B. subtilis by plating on lysogeny broth (LB) agar with 100 μg/mL spectinomycin. Colonies were screened for integration at amyE by colony PCR. All stocks of bacterial strains were stored in 15% glycerol at −80°C.
B. subtilis strains were inoculated from single colonies into 50 mL LB and grown with aeration at 37°C in a shaking incubator (Gyromax 737R, Amerex Instruments, Inc.) for 5 to 6 h until reaching an optical density at 600 nm (OD600) of 1.0 to 1.5. Each culture was diluted 10-fold in LB and split into 13 25-mL cultures containing 0, 25, 250, or 2500 picomoles of a cyanated corrinoid (Cbl, pCbl, CreCba, and Cbi). These cultures were incubated at 37°C with aeration for 3 to 4 h to a final OD600 of 1.5 to 2.0. The cells were pelleted by centrifugation at 4,000 g for 10 min. Cell pellets were rinsed three times by resuspension in 10 mL phosphate-buffered saline solution pH 7.3 followed by centrifugation. After the final centrifugation, tubes were wrapped in aluminum foil to protect adenosylated corrinoids from exposure to light. To extract intracellular corrinoids, cell pellets were resuspended in 5 mL of 100% methanol by vigorous vortexing for 30 s. Samples were stored at −80°C until the next day. Frozen lysates were heated in an 80°C water bath for 1.5 h, with 15 s of vortexing every 30 min. Methanol concentration of each sample was diluted to 10% by adding 45 mL of water, and cell debris was pelleted by centrifugation at 4,000 g for 10 min. The supernatants were used for the subsequent steps. All of the following steps were carried out in darkened rooms illuminated with red light to preserve light-sensitive adenosylated corrinoid samples. Solid-phase extraction of adenosylated corrinoids with Sep-Pak C18 cartridges (Waters) was performed as previously described (82). Solvents were evaporated in a vacuum concentrator centrifuge (Savant SPD1010, Thermo Scientific) at 45°C and the samples were resuspended in 500 μL deionized water and passed through 0.45 μm pore-size filters (Millex-HV, MilliporeSigma). Corrinoids were analyzed on an Agilent 1200 series high-performance liquid chromatography (HPLC) system equipped with a diode array detector (Agilent Technologies). Samples were injected into an Agilent Zorbax SB-Aq column (5-μm pore size, 4.6 mm × 150 mm). The following HPLC method was used: solvent A, 0.1% formic acid–deionized water; solvent B, 0.1% formic acid–methanol; flow rate of 1 mL/minute at 30°C; 25% to 34% solvent B for 11 min, followed by a linear gradient of 34% to 50% solvent B over 2 min, followed by a linear gradient of 50% to 75% solvent B over 8 min.
Corrinoid dose-response assays of riboswitch reporter strains were set up as follows. Saturated cultures of the riboswitch reporter strain in LB were diluted 200-fold in LB and dispensed into 96-well microtiter plates (Corning Costar Assay Plate 3904) containing a range of concentrations of various corrinoids. The plates were sealed with gas diffusible membranes (Breathe-Easy, Diversified Biotech) and incubated at 37°C for 4 to 5 h in a benchtop heated plate shaker (Southwest Science) at 1,200 revolutions per minute (rpm). GFP fluorescence (excitation/emission/bandwidth = 485/525/10 nm) and absorbance at 600 nm (A600) were measured on a Tecan Infinite M1000 Pro plate reader. The A600 measurements of uninoculated medium and fluorescence measurements of the parental control strains lacking gfp were subtracted from all readings. Data were plotted and analyzed in GraphPad Prism 9.
Molecular models of cobalamin in the base-on and base-off/His-on state in complex with various proteins and RNAs were downloaded from the Protein DataBank (PDB) (Table S2) (124). PDB files were analyzed in UCSF Chimera 1.14 (125). Corrinoid molecular models were aligned with each other by the central cobalt atom and coordinating nitrogen atoms of the corrin ring, using the PDB ID 4GMA Cbl model as a reference. Corrinoid models were aligned within the binding sites of riboswitch structures PDB IDs 4GMA, 4FRN, 4GXY, and 6VMY (51, 54, 55).
B. subtilis strains were streaked from frozen stocks onto LB agar plates and incubated overnight at 37°C for 14 to 18 h. Single colonies were used to inoculate 3 mL liquid starter cultures containing Spizizen minimal medium supplemented with 0.02% d-glucose and 0.2% L-Histidine (SMM) (126). Starter cultures of the metH-expressing strain were supplemented with 1 nM CNCbl to support growth, whereas the metE-expressing strains were cultured in SMM without CNCbl. Starter cultures were incubated overnight shaking (250 rpm, 37°C) for 20 h, reaching cell density of about OD600 = 1.0. Starter cultures were diluted 500-fold by transferring 50 μL of starter culture to 25 mL of SMM. Then 75 μL of the diluted culture were dispensed into wells of a 96-well microtiter plate containing 75 μL of SMM supplemented with 40 nM various corrinoids. Plates were sealed with gas diffusible membranes (Breathe-Easy, Diversified Biotech) and incubated at 37°C on the “high shaking” setting of a BioTek Synergy2 plate reader. Growth kinetics and metE and metH expression were measured by A600 and GFP fluorescence every 15 min for 72 h. Data were plotted and analyzed in GraphPad Prism 9. | true | true | true |
PMC9600664 | Xiaofan Zhang,Wenci Gong,Chaohui Duan,Huixia Cai,Yujuan Shen,Jianping Cao | Echinococcus granulosus Protoscoleces-Derived Exosome-like Vesicles and Egr-miR-277a-3p Promote Dendritic Cell Maturation and Differentiation | 14-10-2022 | Echinococcus granulosus,exosome-like vesicles,microRNA,egr-miR-277a-3p,dendritic cells | Cystic echinococcosis, a major parasitic disease caused by Echinococcus granulosus, seriously threatens human health. The excretory–secretory (ES) products of E. granulosus can induce immune tolerance in dendritic cells (DCs) to downregulate the host’s immune response; however, the effect of exosomes in the ES products on the DCs has remained unclear. This study showed that E. granulosus protoscoleces-derived exosome-like vesicles (PSC-ELVs) could be internalized by bone marrow-derived dendritic cells (BMDCs), allowing for the delivery of the parasite microRNAs to the BMDCs. Moreover, PSC-ELVs induced BMDCs to produce the proinflammatory cytokinesinterleukin (IL)-6, IL-12, IL-β, tumor necrosis factor-alpha (TNF-α), and interferon-gamma (IFN-γ). PSC-ELVs also upregulated the BMDCs surface marker major histocompatibility complex class II (MHC II), as well as costimulatory molecules CD40, CD80, and CD86. PSC-ELV-derived egr-miR-277a-3p upregulated the IL-6, IL-12, and TNF-α mRNA levels in BMDCs. Moreover, egr-miR-277a-3p directly targeted Nfkb1 (encoding nuclear factor kappa B 1) to significantly suppress the mRNA and protein levels of NF-κB1 in BMDCs, while the expression of NF-κB p65 significantly increased, suggesting that egr-miR-277a-3p induces the production of proinflammatory cytokines by the modification of the NF-kB p65/p50 ratio in BMDCs. These results demonstrated that PSC-ELVs and egr-miR-277a-3p might enhance DCs maturation and differentiation in a cross-species manner, which in turn may modulate the host immune responses and offer a new approach to echinococcosis prevention and treatment. | Echinococcus granulosus Protoscoleces-Derived Exosome-like Vesicles and Egr-miR-277a-3p Promote Dendritic Cell Maturation and Differentiation
Cystic echinococcosis, a major parasitic disease caused by Echinococcus granulosus, seriously threatens human health. The excretory–secretory (ES) products of E. granulosus can induce immune tolerance in dendritic cells (DCs) to downregulate the host’s immune response; however, the effect of exosomes in the ES products on the DCs has remained unclear. This study showed that E. granulosus protoscoleces-derived exosome-like vesicles (PSC-ELVs) could be internalized by bone marrow-derived dendritic cells (BMDCs), allowing for the delivery of the parasite microRNAs to the BMDCs. Moreover, PSC-ELVs induced BMDCs to produce the proinflammatory cytokinesinterleukin (IL)-6, IL-12, IL-β, tumor necrosis factor-alpha (TNF-α), and interferon-gamma (IFN-γ). PSC-ELVs also upregulated the BMDCs surface marker major histocompatibility complex class II (MHC II), as well as costimulatory molecules CD40, CD80, and CD86. PSC-ELV-derived egr-miR-277a-3p upregulated the IL-6, IL-12, and TNF-α mRNA levels in BMDCs. Moreover, egr-miR-277a-3p directly targeted Nfkb1 (encoding nuclear factor kappa B 1) to significantly suppress the mRNA and protein levels of NF-κB1 in BMDCs, while the expression of NF-κB p65 significantly increased, suggesting that egr-miR-277a-3p induces the production of proinflammatory cytokines by the modification of the NF-kB p65/p50 ratio in BMDCs. These results demonstrated that PSC-ELVs and egr-miR-277a-3p might enhance DCs maturation and differentiation in a cross-species manner, which in turn may modulate the host immune responses and offer a new approach to echinococcosis prevention and treatment.
Cystic echinococcosis is a serious parasitic disease caused by Echinococcus granulosus larvae, which mainly parasitize the liver and lungs of animals and humans [1]. E. granulosus has evolved complex strategies to escape host immune responses, primarily by directing and manipulating the immune response towards tolerance or anergy. Recent research has demonstrated numerous immunoregulatory mechanisms related to macrophages, dendritic cells (DCs), and regulatory T cells (Tregs) [2,3,4]. The early immune response of E. granulosus infection shows Th1-oriented Th2-type responses [5]. E. granulosus oncospheres and cysts could modulate DCs maturation and alter monocyte differentiation to escape the host’s immunosurveillance and promote chronic infection [6]. Prior research by the study investigators showed that the numbers of macrophages, DCs, Tregs, and myeloid-derived suppressor cells (MDSCs) increased in BALB/c mice infected by E. granulosus [2,4], and the expression of arginase-1 (ARG-1) in many kinds of myeloid cells was enhanced, which inhibited the T cell response to E. granulosus antigens [4]. It was also previously found that the E. granulosus excretory–secretory (ES) products could induce immune tolerance in DCs to impair the host’s immune response [7]; however, whether some components of the ES products such as microRNAs (miRNAs) could promote the differentiation and maturation of DCs remains unclear. Excretory–secretory products play a vital role in parasite–host interactions and can downregulate the host’s immune response to mediate immunosuppression and help parasites establish infection [8]. Exosomes are a key component of ES products, comprising a type of membranous vesicles approximately 30–150 nm in diameter, which are released into the extracellular matrix following the fusion of intracellular multivesicular bodies with the cell membrane [9]. Parasite-derived exosomes are involved in the exchange of information among parasites, as well as in the interaction between parasites and their hosts. Parasite-derived exosomes can transmit differentiation, virulence, and drug-resistance genes among parasites [10]. They can also regulate the host’s gene expression and immune responses, in addition to participating in parasite pathogenesis [10]. Exosomes can be effectively internalized by the multiple cells of host, and the biologically active substances they carry such as proteins, lipids, and nucleic acids can then act on those cells, thereby regulating immune responses. Exosomes are rich in noncoding RNAs, especially miRNAs, which can specifically bind to target mRNAs, thus causing mRNA degradation or the inhibition of protein translation and resulting in the downregulation of target protein production. Strikingly, parasite and even plant miRNAs have been detected in animal bodies and human fluids and observed to regulate host genes in a cross-species manner [11]. For example, Schistosoma miR-2162 and miR-1 promote host liver fibrosis by directly targeting the host transforming growth factor beta (TGF-β) receptor III and frizzled-related protein 1, respectively [12,13]. Parasite-derived exosomes can carry a large amount of parasite miRNA, which is involved in the interaction between the parasite and the host, and play a vital role in parasite invasion and infection, pathogenesis, and immune evasion. Primary immune responses are initiated by DCs, which can powerfully activate naive T cells and determine the direction of the body’s immune responses. The process of parasitic infection can induce the production of Tregs and regulatory DCs and inhibit the proliferation and differentiation of T cells [14]. Exosomes from different sources can carry a variety of biological substances, which DCs can internalize to mediate the immune response. Tumor-derived exosomes can influence the differentiation, maturation, and functions of DCs. For example, colon cancer-derived exosomal miR-155 promotes the differentiation and maturation of DCs, which then promote lymphocyte proliferation and induce Th1-type immune responses [15]. Moreover, miR-142 and Let-7i-modified breast cancer exosomes also promote DCs differentiation and maturation [16]. Both E. granulosus protoscoleces (PSCs) and hydatid cysts can release exosome-like vesicles [17,18]. However, the role and mechanism of E. granulosus exosome-derived miRNAs in the differentiation and maturation of DCs remain unclear. Herein, the proportion of DCs in the liver leukocytes of E. granulosus-infected mice was found to be significantly increased at 6-, 9-, and 12-months post-infection. E. granulosus PSC-derived exosomes-like vesicles (PSC-ELVs) could be internalized by mouse bone marrow-derived dendritic cells (BMDCs), allowing the cargo parasite miRNAs to be transported into the BMDCs. Further studies showed that PSC-ELVs promoted the differentiation and maturation of BMDCs, which might help to induce a Th1-type immune response. Moreover, PSC-ELV-derived egr-miR-277a-3p could upregulate the mRNA levels of proinflammatory cytokines (interleukin (IL)-6, IL-12, and tumor necrosis factor-alpha (TNF-α)) in BMDCs. In addition, egr-miR-277a-3p significantly suppressed the mRNA and protein levels of nuclear factor kappa B1 (NF-κB1) in BMDCs by directly targeting the 3′ untranslated region (UTR) of Nfkb1, while the mRNA and protein levels of NF-κB P65 were significantly increased, which suggests that egr-miR-277a-3p might induce the production of proinflammatory cytokines by the modification of the NF-kB p65/p50 ratio in BMDCs.
Female BALB/c mice (6–8 weeks old) were bought from the SLAC Laboratory (Shanghai, China). E. granulosus PSCs were collected from fertile hydatid cysts of naturally infected sheep in the Xinjiang Uygur Autonomous Region, China. PSCs were washed and the viability determined according to the methods of our previous study [19]. BALB/c mice (n = 30) in the infection group were intraperitoneally inoculated with 2000 live PSCs within 200 μL of sterile 0.9% sodium chloride (NaCl), and the control group mice were inoculated with same volume of NaCl. All mice received standard laboratory food and water and were raised in specific pathogen-free conditions.
PSCs (n = 20,000) were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium (Gibco/Life Technologies, Grand Island, NY, USA) supplemented with glucose (4 mg/mL) and antibiotics (200 μg/mL of streptomycin and 200 U/mL of penicillin (Invitrogen, Waltham, MA, USA)) in 75-cm2 cell culture flasks in an incubator (37 °C, 5% CO2). The trypan blue dye exclusion assay was applied to detect the viability of PSCs. We harvested and changed the culture medium of PSCs every 12 h. A total of 300 mL of culture supernatant was collected when PSCs were cultured in serum-free medium for 7 days and then was stored at −80 °C until use.
PSC-ELVs were isolated from the PSCs culture supernatant and enriched with differential centrifugation as our previous study described [19]. Transmission electron microscopy (TEM, HITACHI, Tokyo, Japan) and NanoSight LM10 (NanoSight, Malvern Panalytical Ltd., Malvern, UK) analyses were used to detect the morphology and size distribution of the PSC-ELVs, as previously described [19].
The total protein concentration of the PSC-ELVs was detected with a bicinchoninic acid (BCA) protein measurement kit (Takara, Shiga, Japan). Before measurement, reagents A and B of BCA were mixed thoroughly at a ratio of 100:1 to prepare the working solution. The bovine serum albumin (BSA) standard solution was diluted with deionized water to 0, 125, 250, 500, 750, 1000, 1500, and 2000 μg/mL. Next, 25 μL of each diluted BSA standard solution and PSC-ELVs was added to a 96-well plate and 200 μL of the working solution was then added to each well and mixed immediately. The plate was put in the incubator at 37 °C for 30 min and then cooled to room temperature. The optical density of each concentration of BSA standard solution and PSC-ELVs was detected at 562 nm with a Microplate Reader (Thermo Fisher Scientific, Waltham, MA, USA). The standard curve of the BSA standard solution was drawn by subtracting the average absorbance value of the blank well (Figure S1) and then using that value to calculate the protein concentration of the PSC-ELVs.
BMDCs were prepared according to the methods of a previous study, with minor modifications [20]. Briefly, bone marrow cells were isolated from the leg bones of wild-type BALB/c mice and then cultured in RPMI 1640 supplemented with 10% fetal bovine serum (Gibco), 2 ng/mL of IL-4 (PeproTech, Cranbury, NJ, USA) and 20 ng/mL of granulocyte-macrophage colony-stimulating factor (PeproTech) and maintained in an incubator (37 °C, 5% CO2). The medium was renewed after the initial bone marrow cells were cultured for 4 days. The cells were collected on the 7th day, and the purity of the CD11b+CD11c+ cells (BMDCs) was >95%, as detected with flow cytometry. BMDCs were treated as immature DCs in the experiments. For stimulation, 2.5 μg/mL of PSC-ELVs was added to the BMDCs after culture for 7 days, and the cells were then collected at different times according to the experimental requirements. BMDCs were transfected with 20 μΜ of egr-miR-277a-3p mimic (RiboBio, Guangzhou, China), 20 μΜ of egr-miR-277a-3p inhibitor (RiboBio), or negative controls (NC, RiboBio), according to the manufacturer’s protocols.
To investigate whether the PSC-ELVs could be internalized by the host BMDCs, PSC-ELVs were stained with the fluorescent lipid dye PKH67 (Sigma-Aldrich, St. Louis, MO, USA) according to the methods of a previous study with minor modifications [21]. Briefly, the PKH67-labeled PSC-ELVs (5 μg) were washed with phosphate-buffered saline (PBS) and concentrated by ultracentrifugation at 110,000× g for 90 min to remove the non-absorbed lipid dye. BMDCs were seeded in a 6-well plate and then co-incubated with the labeled PSC-ELVs (2.5 μg/mL) for 10 min. Meanwhile, an equal volume of PBS was co-incubated with the BMDCs as a control group. Following a 10min incubation, the culture medium was removed and BMDCs washed twice with PBS. Then, 4% paraformaldehyde was added to fix the cells for 20 min. The cell nuclei were stained with 4′6-diamidino-2-phenylindole (DAPI, Beyotime Biotechnology, Jiangsu, China). Finally, the internalization of the PSC-ELVs by the host BMDCs was observed under a confocal fluorescence microscope (Nikon, Tokyo, Japan).
The total small RNA in PSC-ELVs was extracted with an miRNeasy Serum/Plasma Kit (Qiagen, Hilden, Germany) and immediately reversed transcribed into cDNA with a miScript II RT Kit (Qiagen). The levels of parasite miRNA were detected using quantitative real-time PCR (qPCR) employing a QuantiFast SYBR Green PCR Kit (Qiagen). The primers used for the qPCR step of the qRT-PCR protocol are listed in Supplemental Table S1. U6 snRNA was used as an endogenous control to normalize the qPCR data, and the relative levels of miRNA were analyzed with the 2–ΔΔCt method [22]. The total RNA in BMDCs was extracted with TRIzol reagent (Invitrogen) and then reversed transcribed into cDNA using a PrimeScript™ RT reagent Kit with gDNA Eraser (Takara). The cDNA was used as a template in a qPCR reaction system containing TB Green Premix Ex Taq (Takara) with 0.4 μΜ of forward and reverse primers. The primers used for the qPCR detection of mRNA are listed in Supplemental Table S2. The relative gene expression values were normalized to that of Gapdh (encoding glyceraldehyde-3-phosphate dehydrogenase) and analyzed with the 2–ΔΔCt method.
The treated BMDCs were washed twice with precooled PBS and then lysed with a Radioimmunoprecipitation (RIPA) lysis solution containing phosphatase and protease inhibitors on ice for 30 min. The cell lysates were separated using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to a polyvinylidene difluoride (PVDF) membrane (Merck Millipore, Darmstadt, Germany). After blocking nonspecific binding sites, the membranes were incubated with different primary antibodies [anti-14-3-3 zeta/delta, anti-enolase-1, anti-CD9, anti-NF-κB1, anti-NF-κB p65, and anti-GAPDH (Cell Signaling Technology, MA, USA)] and their respective horseradish peroxidase (HRP)-conjugated secondary antibodies. The membranes were visualized by an chemiluminescence (ECL) detection system (Merck Millipore). The intensities of the immunoreactive protein bands were analyzed with Image J (NIH, Bethesda, MD, USA).
BMDCs were stimulated with PSC-ELVs for 12 h. Then, the protein levels of IL-6, IL-12p70, IL-1β, TNF-α, interferon-gamma (IFN-γ), and IL-10 were detected in the supernatants of BMDCs using ELISAs (eBioscience, San Diego, CA, USA) following the manufacturer’s instructions.
Flow cytometry was used to detect the expression of major histocompatibility complex class II (MHC II) and costimulatory molecules on the surface of BMDCs stimulated by PSC-ELVs. The treated BMDCs were incubated with the following fluorescently labeled antibodies at 4 °C for 30 min: BV421-labeled anti-CD11b, FITC-labeled anti-CD11c, PE-labeled anti-MHC II, APC-labeled anti-CD86, PE-Cy7-labeled anti-CD80, and PE-Cy7-labeled anti-CD40 (BioLegend, San Diego, CA, USA). An LSRFortessa X-20 instrument (BD Biosciences, Franklin Lakes, NJ, USA) and FlowJo software (Tree Star Inc., Ashland, OR, USA) were used to acquire and analyze the flow cytometry data.
According to the online miRDB database, the seed sequence of egr-miR-277a-3p was predicted to be complementary with the 3′ UTR of Nfkb1. Wild-type and mutant 3′ UTRs of Nfkb1 containing the predicted egr-miR-277a-3p binding sites were chemically synthesized and then cloned into the pmirGLO vector (Promega, Madison, WI, USA). Seeded in a 96-well plate, the 293T cells were transfected with 50 nM of egr-miR-277a-3p mimics or a negative control (RiboBio) and were co-transfected with 250 ng of wild-type or mutant Nfkb1 3ʹ UTR plasmid, using Lipofectamine 2000 (Invitrogen) following the manufacturer’s protocols. The cells were collected after transfection for 48 h, and the luciferase activity of the Nfkb1 3ʹ UTR plasmid was detected with a Dual-Glo Luciferase Assay System (Promega).
Data are presented as the mean ± SD and were analyzed with an unpaired Student’s t-test or one-way analysis of variance (ANOVA) using GraphPad Prism version 7.0 (GraphPad Software Inc., San Diego, CA, USA). Differences were considered statistically significant when the p-value was less than 0.05.
Transmission electron microscopy (TEM) and nanoparticle tracking analysis (NTA) were performed to evaluate the morphology and size distribution of the PSC-ELVs. Rounded or cup-shaped vesicles with a diameter of 30–150 nm were observed under TEM (Figure 1a). NTA also demonstrated that the most purified vesicles isolated from PSCs were between 60–80 nm in diameter (Figure 1b), which was consistent with the particle size of exosomes [23]. Western blotting analysis further confirmed that the exosomal marker proteins, including CD9, enolase-1, and 14-3-3 zeta/delta, were present in the PSC-ELVs (Figure 1c).
Exosomes play a vital role in host–parasite interactions. As such, the current study investigators previously isolated PSC-ELVs and identified the ncRNA profiles that might be involved in the regulation of host immunity [19]. Egr-miR-277a-3p was found to be one of the most abundant parasite miRNAs in PSC-ELVs. To confirm whether E. granulosus PSC-ELVs can be internalized by host cells, the PSC-ELVs were labeled with the lipid dye PKH67, and the labeled exosomes were incubated with BMDCs in vitro. The results showed that the PSC-ELVs could be internalized by BMDCs (Figure 2a). To further observe whether PSC-ELVs can transport the E. granulosus miRNA cargo into BMDCs, qRT-PCR was performed to detect the levels of the 10 most abundant miRNAs found in PSC-ELVs in BMDCs. After being incubated with PSC-ELVs for 4 h, the levels of the 10 most abundant miRNAs (egr-miR-277a-3p, egr-bantam-3p, egr-let-7-5p, egr-miR-4989-3p, egr-miR-10a-5p, egr-miR-2162-3p, egr-miR-125-5p, egr-miR-71-5p, egr-miR-61-3p, and egr-miR-2a-3p) in BMDCs were found to be significantly increased (Figure 2b). These data suggested that PSC-ELVs can be effectively internalized by BMDCs, meanwhile, E. granulosus miRNAs such as egr-miR-277a-3p can be transferred into BMDCs.
DCs are powerful antigen-presenting cells that can effectively initiate the body’s primary immune response and further determine the direction of the body’s immune response during a parasitic infection. To profile the dynamic changes of DCs, flow cytometry was performed to detect the ratio of DCs in the liver leukocytes of mice after infection with E. granulosus (at 3, 6, 9, and 12 months). Compared with the control group, the level of DCs in liver leukocytes was increased at 6, 9, and 12 months post-infection (F(3, 22) = 201.4, p < 0.0001) and exhibited an upward trend for the duration of the infection (Figure 3a,b). To further study the role of E. granulosus PSC-ELVs in the differentiation and maturation of BMDCs, the expression levels of cytokines and surface molecules of BMDCs were detected after incubation with PSC-ELVs. BMDCs were stimulated with PSC-ELVs (2.5 μg/mL) and lipopolysaccharide (LPS, 100 ng/mL). After 4 h, the BMDCs were collected to extract total cellular RNA, and then qRT-PCR was performed to determinate the mRNA levels of cytokines in the BMDCs. The results showed that compared with the blank control group, PSC-ELVs could effectively upregulate the mRNA levels of proinflammatory factors IL-6, IL-12, TNF-α, IL-1β, and IFN-γ and inducible nitric oxide synthase (iNOS), as well as anti-inflammatory factors IL-10 and transforming growth factor beta (TGF-β) in BMDCs, while they downregulated the mRNA levels of anti-inflammatory factor indoleamine 2,3-dioxygenase 1 (IDO) (Figure 4a). In general, the mRNA levels of proinflammatory factors increased significantly and showed clear proinflammatory characteristics. Compared with the LPS-only group, co-stimulation with PSC-ELVs and LPS had a synergistic effect, inducing BMDCs to increase the mRNA levels of IL-6, IL-12, IL-1β, iNOS, IFN-γ, and IL-10 (Figure 4a). Moreover, ELISA was used to further determine the secretion of cytokines in the supernatant of BMDCs after incubation with PSC-ELVs or LPS for 24 hours. The ELISA results suggested that compared with the blank control group, PSC-ELVs could effectively induce BMDCs to produce high levels of the proinflammatory factors IL-6, IL-12, TNF-α, IL-1β, and IFN-γ, as well as low levels of the anti-inflammatory factor IL-10 (Figure 4b). Compared with the LPS-only group, co-stimulation by PSC-ELVs and LPS induced higher levels of IL-6, IL-12, TNF-α, IL-1β, IFN- γ, and IL-10, thus exhibiting a synergistic effect (Figure 4b). This result was consistent with the results of qRT-PCR. Overall, PSC-ELVs caused BMDCs to produce an inflammatory immune response, inducing a stimulatory DCs polarized phenotype in vitro and demonstrating a synergistic effect with LPS. DCs are a type of professional antigen-presenting cell, and their surface MHC Ⅱ and costimulatory molecules play a vital role in the process of antigen presentation. To explore whether PSC-ELVs affect the expression levels of BMDCs surface MHC II and costimulatory molecules, flow cytometry was applied to detect the levels of BMDCs surface markers after stimulation with 2.5 μg/mL PSC-ELVs or 100 ng/mL LPS for 24 h. PSC-ELVs were found to effectively upregulate the expression of MHC II molecules as well as that of CD40, CD80, and CD86 on BMDCs (Figure 5a,b). Compared with the LPS-only group, co-stimulation by PSC-ELVs and LPS downregulated the expression of MHC II and CD86 but caused no significant changes in the other co-stimulatory molecules (Figure 5). Thus, these results suggested that PSC-ELVs promote the differentiation and maturation of BMDCs.
Egr-miR-277a-3p is one of the most abundant miRNAs in PSC-ELVs and can be transferred into BMDCs via PSC-ELVs. To study the role of egr-miR-277a-3p in PSC-ELV-mediated induction of proinflammatory cytokine production in BMDCs, an egr-miR-277a-3p mimic or inhibitor was transfected into BMDCs, and the cytokines mRNA levels were detected using qRT-PCR. Egr-miR-277a-3p is a parasite-derived miRNA; therefore, PSC-ELVs were added to provide a target for the egr-miR-277a-3p inhibitor. The results showed that the egr-miR-277a-3p mimic could upregulate the mRNA levels of IL-6, IL-12, and TNF-α in BMDCs. Meanwhile, the mRNA levels of anti-inflammatory factors IL-10 and ARG-1 were significantly decreased, while IL-1β mRNA levels showed no significant changes (Figure 6a), which showed clear proinflammatory properties overall. By contrast, in BMDCs, the egr-miR-277a-3p inhibitor induced decreased mRNA levels of IL-6 and IL-12, while there was no significant change in the IL-1β mRNA (Figure 6b). This was consistent with the transfection results of the egr-miR-277a-3p mimic, which showed proinflammatory properties. No significant change was seen in the mRNA levels of TNF-α and ARG-1 in BMDCs transfected with the egr-miR-277a-3p inhibitor. However, the egr-miR-277a-3p inhibitor downregulated the mRNA level of IL-10 in BMDCs, which contrasted with the transfection results of the egr-miR-277a-3p mimic. These data proved that egr-miR-277a-3p could induce BMDCs to upregulate the mRNA levels of inflammatory cytokines, which might influence the differentiation and maturation of BMDCs.
The target genes of egr-miR-277a-3p were predicted using the online bioinformatics software miRDB. Among them, Nfkb1 was identified as a potential target gene of egr-miR-277a-3p. Nfkb1 encodes the NF-κB p105 protein, which is processed to generate the p50 subunit of NF-ĸB. qRT-PCR and western blotting were applied to determine the mRNA and protein levels of NF-κB1 in BMDCs after transfection with the egr-miR-277a-3p mimic. qRT-PCR results suggested that the egr-miR-277a-3p mimic could suppress the mRNA levels of Nfkb1 (Figure 7a). At the same time, the mRNA levels of Rela (encoding the p65 subunit of NF-κB) were significantly increased (Figure 7a). In its canonical signaling pathway, NF-κB is a heterodimer consisting of p65 and p50 subunits. The p65 protein functions as the activator of the p65-p50 heterodimer, and an increase in the NF-κB p65/p50 ratio is beneficial to the secretion of inflammatory factors TNF-α and IL-6 [24,25]. Western blotting showed that the NF-κB P65 protein levels were significantly increased after the transfection of the egr-miR-277a-3p mimic into BMDCs, while the NF-κB1 levels were significantly decreased (Figure 7b). Thus, egr-miR-277a-3p markedly increased the NF-κB p65/p50 ratio in BMDCs. To confirm whether Nfkb1 is a direct target of egr-miR-277a-3p, luciferase reporter constructs were generated containing wild-type or mutated Nfkb1 3ʹ UTR (Nfkb1-3ʹ UTR-WT or Nfkb1-3ʹ UTR-Mut) complementary sites of egr-miR-277a-3p (Figure 7c). The 293T cells were co-transfected with the constructs and either egr-miR-277a-3p mimics or the mimic NC. Luciferase activity was found to be significantly decreased in the cells transfected with the Nfkb1-3ʹ UTR-WT and egr-miR-277a-3p mimics compared to the cells transfected with the NC mimic (Figure 7d). Conversely, the mutation of four nucleotides in the seed-binding sequence of the Nfkb1 3ʹ UTR induced a complete abrogation of that inhibition and thus increased luciferase activity (Figure 7d). These data indicated that egr-miR-277a-3p directly targets Nfkb1, which might upregulate the expression of p65 to promote the expression of the inflammatory factors IL-6, IL-12, and TNF-α in BMDCs.
Parasite-derived exosomes play an important role in the exchange of information among parasites and in the interaction between parasites and their hosts. Parasite miRNAs and antigens carried by parasite-derived exosomes play a vital role in these processes. DCs can potently activate naive T cells, initiate primary immune responses, and determine the direction of the body’s immune response during a parasitic infection. E. granulosus ES antigens regulate the maturation and function of DCs, affect the release of inflammatory factors, and regulate Th1/Th2 immune responses [7]. However, the functions and mechanisms of both E. granulosus PSC-ELVs and their abundant miRNAs toward DCs have not yet been reported. We isolated and identified PSC-ELVs and found that E. granulosus miRNAs could be transported into BMDCs to regulate immune responses via the uptake of PSC-ELVs by BMDCs. Many studies have demonstrated that ELVs can be internalized by DCs, transferring their contents, e.g., miRNAs, into the DCs cytosol to regulate immune responses [26,27,28]. Parasite infection initially promotes the production of stimulatory DCs to help the host to kill parasites, while it will gradually induce the production of regulatory DCs to help the parasites to evade immune responses. Immature DCs differentiate into stimulatory DCs when stimulated by proinflammatory signals such as pathogen-associated molecular patterns and secrete proinflammatory factors such as IL-6, IL-12, IFN-γ, and TNF-α. However, under the stimulation of inhibitory signals such as TGF-β and IL-10, they differentiate into regulatory DCs and secrete anti-inflammatory molecules such as TGF-β, IL-10, ARG, and IDO [14]. To the best of our knowledge, this is the first study to show that PSC-ELVs can induce immature DCs to differentiate into stimulatory DCs. E. granulosus PSC-ELVs were found to induce BMDCs to produce high levels of IL-6, IL-12, TNF-α, IL-1β, and IFN-γ and low levels of IL-10, which are clear proinflammatory characteristics. In addition, co-stimulation with PSC-ELVs and LPS had a synergistic effect, inducing BMDCs to produce higher levels of these cytokines. Furthermore, PSC-ELVs induced BMDCs to upregulate the expressions of MHCII, CD40, CD80, and CD86 molecules on their surface, which helped to promote BMDCs maturation. The percentage of DCs was found to increase in the liver leukocytes of mice infected with E. granulosus. Moreover, a previous related study demonstrated that these DCs highly express ARG-1, which inhibited the T cell response to E. granulosus antigens [4]. Therefore, E. granulosus PSC-ELVs induce the production of stimulatory DCs, which might serve as a new strategy to enhance the anti-parasitic immunity. However, the immune functions of E. granulosus PSC-ELVs and PSC-derived ES are likely different. E. granulosus adult worm antigen (AWA) and adult ES, both derived from the adult stages of E. granulosus, trigger different immune responses [7]. Research has shown that PSC-derived ES can significantly inhibit proinflammatory responses by directly inducing B10 cells and inhibiting B17 and Th17 cells, thereby downregulating anti-parasitic immunity [29,30]. Therefore, in the current study, it was speculated that PSC-ELVs may have different immune effects and might carry E. granulosus-related virulence factors that promote the maturation and differentiation of DCs, thus having the potential to induce Th1-type immune responses and even enhance host immunity. NF-κB1 is necessary for DCs to promote optimal Th2-type responses, which might have therapeutic potential for helminthiasis dominated by Th2 immune responses [31]. Furthermore, our study found that egr-miR-277a-3p in PSC-ELVs might induce the polarization of BMDCs toward stimulatory DCs by directly targeting Nfkb1. egr-miR-277a-3p significantly upregulated the mRNA levels of proinflammatory factors IL-6, IL-12, and TNF-α and downregulated the anti-inflammatory factors IL-10 and ARG-1 in BMDCs, which might help BMDCs differentiate into stimulatory DCs. Nfkb1 was found to be a potential direct target gene of egr-miR-277a-3p using qRT-PCR, western blotting, and Luciferase reporter assays. Meanwhile, in the current study, it was found that egr-miR-277a-3p could significantly upregulate the expression of NF-κB p65 in BMDCs. NF-κB can mediate inflammatory responses, and an increase in p65/p50 expression promotes the secretion of IL-6 and TNF-α [24,32]. Therefore, it was speculated that egr-miR-277a-3p could directly target Nfkb1 and might induce the production of these proinflammatory cytokines by modification of the NF-kB p65/p50 ratio in BMDCs. As such, the blockade of NF-κB1 might prove valuable in DC-based therapies for echinococcosis. This study did have some limitations, as it did not explore whether egr-miR-277a-3p could promote the maturation and differentiation of DCs in vivo. Therefore, the injection of an egr-miR-277a-3p agomir into E. granulosus-infected mice or NF-κB1 deficient mice to confirm the functions of egr-miR-277a-3p will be included in a future study.
In conclusion, this study found that PSC-ELVs could promote BMDCs maturation and differentiation into stimulatory DCs, which is beneficial to mediating a Th1 immune response. Further, egr-miR-277a-3p could directly target Nfkb1 to suppress its expression, which induced DCs to produce proinflammatory cytokines in a cross-species manner. Therefore, PSC-ELVs and egr-miR-277a-3p could regulate the host immune response, thus providing a new strategy to prevent and treat echinococcosis. | true | true | true |
|
PMC9600678 | Xuan-Cheng He,Jian Wang,Hong-Zhen Du,Chang-Mei Liu,Zhao-Qian Teng | Correction: He et al. Intranasal Administration of Agomir-let-7i Improves Cognitive Function in Mice with Traumatic Brain Injury. Cells 2022, 11, 1348 | 17-10-2022 | Correction: He et al. Intranasal Administration of Agomir-let-7i Improves Cognitive Function in Mice with Traumatic Brain Injury. Cells 2022, 11, 1348
The authors wish to make the following changes to their paper [1].
In the text of this article initially published (PDF version), there were typing errors in the last paragraph of Page 10. The text originally reading ‘Next, we examined whether there are any expression changes in STING in agomir-let-7i-treated HSI brains. Consistent with our expectations, a reduced STING mRNA expression level was observed in the hippocampi of agomir-let-7i-treated HSI mice compared to scramble-treated HSI mice (F(2,6) = 169.3, p < 0.001; Sham vs. HSI + Scramble, p < 0.001; Sham vs. HSI + Agomir-let-7i, p = 0.803; HSI + Scramble vs. HSI + Agomir-let-7i, p < 0.001) (Figure 6D).’ has been amended to read ‘Next, we examined whether there are any expression changes in STING in agomir-let-7i-treated uninjured brains. Consistent with our expectations, a reduced STING mRNA expression level was observed in the hippocampi of agomir-let-7i-treated mice compared to scramble-treated mice (F(2,6) = 169.3, p < 0.001; Vehicle vs. Agomir-let-7i, p < 0.001; Vehicle vs. Scramble, p = 0.803; Scramble vs. Agomir-let-7i, p < 0.001) (Figure 6D)’.
Figures 3, 4 and 6 should be replaced. In the original publication, there were labeling errors in Figures 3 and 6. In Figure 3B, the labels ‘HIS + Scramble’ and ‘HIS + Agomir-let-7i’ were inadvertently swapped. The corrected Figure 3 appears below. In Figure 4A of the initially published article, the separate image of TUNEL staining in the HIS + Agomir-let-7i group was erroneously flipped horizontally. The error does not affect the conclusions reported in the paper. The corrected Figure 4 appears below. In Figure 6D, the x-axis label, now reading ‘Vehicle’ and ‘Scramble’, initially appeared as ‘‘Control’ and ‘Agomir-NC’. In Figure 6E, the x-axis label, now reading ‘HIS + Scramble’ and ‘HIS + Agomir-let-7i’, initially appeared as ‘Stroke + Scramble’ and ‘Stroke + Agomir-let-7i’. In the Figure 6D caption, the text originally reading ‘the hippocampi of sham, scramble-treated and agomir-let-7i-treated HSI mice’ has been amended to read ‘the hippocampi of Vehicle-, scramble- and agomir-let-7i-treated uninjured mice’. The corrected Figure 6 appears below. The authors apologize for any inconvenience caused and state that the scientific conclusions are unaffected. This correction was approved by the Academic Editor. The original publication has also been updated. | true | true | true |
|||
PMC9600689 | Junwei Li,Yuexia Wang,Bin Wang,Juan Lou,Peng Ni,Yuefei Jin,Shuaiyin Chen,Guangcai Duan,Rongguang Zhang | Application of CRISPR/Cas Systems in the Nucleic Acid Detection of Infectious Diseases | 11-10-2022 | CRISPR/Cas system,biosensing technologies,pathogen nucleic acids,SARS-CoV-2,diagnoses | The CRISPR/Cas system is a protective adaptive immune system against attacks from foreign mobile genetic elements. Since the discovery of the excellent target-specific sequence recognition ability of the CRISPR/Cas system, the CRISPR/Cas system has shown excellent performance in the development of pathogen nucleic-acid-detection technology. In combination with various biosensing technologies, researchers have made many rapid, convenient, and feasible innovations in pathogen nucleic-acid-detection technology. With an in-depth understanding and development of the CRISPR/Cas system, it is no longer limited to CRISPR/Cas9, CRISPR/Cas12, and other systems that had been widely used in the past; other CRISPR/Cas families are designed for nucleic acid detection. We summarized the application of CRISPR/Cas-related technology in infectious-disease detection and its development in SARS-CoV-2 detection. | Application of CRISPR/Cas Systems in the Nucleic Acid Detection of Infectious Diseases
The CRISPR/Cas system is a protective adaptive immune system against attacks from foreign mobile genetic elements. Since the discovery of the excellent target-specific sequence recognition ability of the CRISPR/Cas system, the CRISPR/Cas system has shown excellent performance in the development of pathogen nucleic-acid-detection technology. In combination with various biosensing technologies, researchers have made many rapid, convenient, and feasible innovations in pathogen nucleic-acid-detection technology. With an in-depth understanding and development of the CRISPR/Cas system, it is no longer limited to CRISPR/Cas9, CRISPR/Cas12, and other systems that had been widely used in the past; other CRISPR/Cas families are designed for nucleic acid detection. We summarized the application of CRISPR/Cas-related technology in infectious-disease detection and its development in SARS-CoV-2 detection.
Infectious diseases, causing a massive burden of disability and death, have always been terrifying threats to humans [1]. Zika virus [2], Ebola virus [3], severe acute respiratory syndrome coronavirus (SARS-CoV), and Middle East respiratory syndrome coronavirus (MERS-CoV) [4], as well as the ongoing outbreak of the SARS-CoV-2 epidemic, has caused great harm to human beings [5]. Fast, accurate, and cost-effective detection and identification of pathogens are key to the control and management of infectious-disease epidemics. According to the World Health Organization, the ideal pathogen-detection method should be fast, specific, sensitive, and not require too-large equipment. The traditional method for pathogen detection was plate culture, which often takes a long time and effort. Following the emergence of real-time polymerase chain reaction (PCR), people can accurately identify pathogenic pathogens through pathogen-specific biomarkers and greatly shorten the detection time. However, real-time PCR also has some limitations, such as requiring complex infrastructure and skilled operators, which limits its deployment in countries with relatively backward medical standards. The clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR- associated protein (Cas) system [6] is a sensitive and specific biological system that can quickly identify pathogen-specific nucleic acids [7]. CRISPR systems are present in approximately 90% of archaea and 48% of bacterial genomes and are essential for their genetic components [8,9]. They play an important role in immune function and protect the host against foreign genes [10]. In CRISPR systems, foreign mobile genetic elements (MGE) are firstly cleaved by synthesized corresponding proteins to generate a spacer, then integrated into the CRISPR array; this is the acquisition stage; the second stage is CRISPR RNA (crRNA) biosynthesis. The CRISPR array transcribes and processes the biosynthesis of mature crRNA [11]; the last is the interference stage; the invading MGEs are degraded after being recognized by crRNA [12]. CRISPR systems also show a corresponding diversification in both their Cas protein sequence and species [13]. According to the design principles of the effector module, all CRISPR systems fall into two distinct categories: class 1 systems utilize a large multi-Cas protein complex for crRNA binding and target sequence degradation, while class 2 systems employ a single DNA endonuclease to recognize double-stranded DNA (dsDNA) substrates and cleave each strand with a distinct nuclease domain (HNH or RuvC) [13,14]. CRISPR systems brightly released their potential in various field diagnosis and genotyping applications and brought a leap forward developing opportunities in the field of pathogen detection [15,16]. Some CRISPR systems have been developed to detect nucleic acids or pathogen biomarkers. CRISPR-based biosensing platforms are expected to revolutionize the diagnosis of pathogens [17,18]. The field of CRISPR system exploration is growing rapidly. More and more CRISPR systems are being discovered and applied in nucleic acid detection [19,20]. In this paper, we focused on the more comprehensive application of CRISPR systems in biosensing platforms and showed examples of the detection of the nucleic acids of various pathogens.
CRISPR systems are a universal immune mechanism. Due to this universal adaptability, CRISPR systems are indeed as diverse as the innate immune system. The Cas protein sequence and genome organization of the CRISPR-Cas locus show great diversity. All CRISPR systems are divided into two different categories [21]. Class 1 systems have multi-subunit effector complexes containing multiple Cas proteins, while in class 2 systems, the effector is a single large multi-domain protein. The two CRISPR-Cas types are further divided into six types. Class I includes type I, III, and IV, and class II comprises type II, V, and VI. Each type is characterized by a different structure of effector modules. The Cas effector can be a single Cas protein or a large multi-Cas protein complex; though each established CRISPR-based nucleic acid biosensing system has a different combination of components, the fundamental difference is that they use different Cas effectors. For example, Cas9 contains two nuclease domains, RuvC and HNH, which cleave the target and non-target strands of dsDNA, respectively. CRISPR/Cas12, CRISPR/Cas13, and CRISPR/Cas14 have collateral strand cleavage activity, that is, non-specifically cleaving the surrounding nucleic acid molecules after recognizing the target sequence, and CRISPR/Cas10 will trigger its nuclease activity after binding to a perfectly matched target sequence. At present, the most widely used CRISPR systems, such as CRISPR/Cas9, CRISPR/Cas12, CRISPR/Cas13, and CRISPR/Cas14, belong to the type II CRISPR system, and their effector template is a single protein. These systems are easy to be edited for nucleic acid detection [22]. Recently, many researchers have conducted in-depth research on CRISPR/Cas10, which belongs to the type III CRISPR system, and developed some high-sensitivity detection platforms. CRISPR/Cas10 uses RNA as a target sequence for identification and performs well in the diagnosis of SARS-CoV-2 [23]. The reported CRISPR-systems-based biosensing system can be divided into five categories (CRISPR/Cas9, CRISPR/Cas12, CRISPR/Cas13, CRISPR/Cas14, and CRISPR/Cas10). Table 1 lists the main features of the various CRISPR-related detection systems.
The CRISPR/Cas9 system (Figure 1A), a typical class 2 type II CRISPR system, is widely applied at present [51]. In 2013, the type II CRISPR system was first isolated from Streptococcus pyogenes (SpCas9) and successfully performed RNA-guided DNA cleavage in mammalian cells, paving the way for the CRISPR/Cas9 system as a widely available genome editing tool [52,53]. Under natural processes, the trans-activating crRNA (tracrRNA) base pairs with the repeat sequence in the crRNA to form a unique dual RNA hybrid structure guide that directs Cas9 to cleave the target DNA, so a chimeric sgRNA was designed that combines crRNA and tracrRNA into a single RNA transcript, simplifying the system while preserving Cas9-mediated full-function sequence-specific DNA cleavage [14,54]. Cas9 contains two nuclease domains, RuvC [55] and HNH [56], which cut the target DNA strands and non-target DNA strands respectively [57]. A short trinucleotide protospacer adjacent motif (PAM) is also essential for initial target sequence recognition [58]; the target sequence could not be recognized without a corresponding PAM site. After successful identification, a double-strand break (DSB) occurs upstream of the 3′-NGG PAM site. Currently, the CRISPR/Cas9 system has been widely used in genome editing, single-nucleotide-mutation detection, and other fields [59,60,61,62].
Nucleic acid sequence-based amplification (NASBA)-CRISPR Cleavage (NASBACC) [24] (Figure 2A), based on the CRISPR/Cas9 system, was designed to detect the Zika virus. First, the reverse primer of NASBA was used to amplify the sample RNA sequence [63], and a specific sensor trigger sequence was added to the amplified product. CRISPR/Cas9 was introduced to achieve a specific cleavage reaction. When the target sequence is present, the dsDNA synthesized by the NASBA reaction will be cleaved by CRISPR/Cas9, and the truncated RNA product will not activate the sensor switch. However, if no target sequence existed, the added sensor sequence can trigger a subsequent reaction to cause color changes and finally achieve detection. In the Cas-exponential amplification reaction (EXPAR) [25] (Figure 2B), the Cas9/sgRNA complex mediates the site-specific cleavage of ssDNA substrates, produce cleaved fragments (X). X hybridizes to the EXPAR template and is extended along the template by DNA polymerase from its 3′ end. The subsequently formed duplex is cleaved by Nease; a copy of X is released from the template, and then the dissociated X continues the same process as above, serving to amplify the signal. At the same time, the mixture was incubated in a real-time PCR assay system, and the fluorescence intensity was monitored at 1 min intervals, with a detection limit of 0.82 Attomolar (aM). Cas-EXPAR could be used for site-specific DNA methylation detection by combining it with bisulfite conversion. There are also CRISPR/Cas9-based detection systems designed with the UiO66 platform [28] (Figure 2C). UiO66 is a nanoporous material [64]. The fluorescence of ssDNA is quenched in the metal–organic framework platform based on UiO66 [28]. Two Cas9/sgRNA ribonucleoprotein complexes were designed to recognize and cleft the target DNA to produce short ssDNA. Then the circular probe hybridized with the short ssDNA to form long repetitive ssDNA by rolling loop amplification [65]. In the presence of long-ssDNA, the fluorescent probe leaves UiO66 and hybridizes with long-ssDNA, leading to fluorescence recovery. Therefore, the quantitative detection of target DNA can be evaluated by the recovered fluorescence intensity. Some platforms achieve nucleic acid detection through the sequence-specific binding ability of inactivated Cas9 effector [66]. Paired dCas9 (PC) reporter system [29] (Figure 3A), designed two paired dCas9/sgRNA ribonucleoprotein complexes that are respectively connected to the N-terminal and C-terminal of the firefly luciferase (NFluc and CFluc) [67]. Upstream and downstream sequence segments are adjacent; when these two fragments are present, two dCas9/sgRNA ribonucleoprotein complexes combined and activated luciferase activity, finally generating luminescence for detection. The VirD2(virulence D2)-dCas9 guided and LFA-coupled nucleic acid test (Vigilant) [27] (Figure 3B), combining CRISPR/Cas9 with lateral flow analysis (LFA) technology, designed a direct-detection platform. The deactivated Cas9 of Streptococcus pyogenes (spdCas9) was fused with VirD2 relaxase [68] and designed as a reporter ssDNA with the 5′-end of the 25-BP VirD2 recognition sequence and the 3′-end of the biotinylated sequence. When the target nucleic acid sequence was presented, the biotin-reporting ssDNA-VirD2-Cas9-sgRNA-targeting ssDNA complexes formed and finally reported results by LFA [69]. Recently, CRISPR/Cas9 has also been integrated into sequencing systems, such as finding low abundance sequences by hybridization–next-generation sequencing (FLASH-NGS) [26] (Figure 3C). In this system, the combination of recombinant Cas9 and multiple-guide RNAs can accurately remove unwanted background sequences [70] and also can target almost any interest sequence without optimization. The mechanism was to seal the sample genomic DNA or cDNA with phosphatase, then digest it through multiple combinations; each sgRNA-guided Cas9/sgRNA ribonucleoprotein complexes can cut the interest sequence into Illumina sequencing-size fragments, finally generating a cleavage product that can be connected to a universal sequencing connector. Through subsequent amplification, the target sequence was enriched and combined with sequencing flow cells to achieve multiple detections. This method shows good sensitivity in highly multiple detections of antibiotic genes and has certain guiding significance in the development of cancer-mutation detection, rare mosaic allele detection, and the targeted transcriptomics of clinical samples.
The detection platform based on CRISPR/Cas9 has brought new opportunities to the detection of infectious diseases, but there are still large deficiencies in the whole system. At first, it can allow up to six consecutive mismatches in the 5′-end region of the prototype gasket, which greatly increases the occurrence of off-target effects. Second, it requires 3′GC-rich PAM for anchoring, which has some limitations in selecting target gene segments for detection. In contrast to RT-PCR, NASBACC uses a paper-based sensor platform to advance field-ready diagnostics, but even when strains of two different lineages are mismatched up to 4-nt (11%), it can also fully activate the sensor, limiting its value in practical applications. Cas-EXPAR does not require exogenous primers for detection, avoiding target-independent amplification, and can be used as a general method for the detection of DNA, RNA, methylated DNA, and other nucleic acids, with high specificity and rapid amplification kinetics, which has great application potential in biological analysis and disease diagnosis. The UiO66-platform-based Cas9 provides significant sensitivity under mild reaction conditions and has good selectivity for different pathogens. It also shows good performance in actual sample analysis, but it is three orders of magnitude lower than the detection limit of RT-PCR. Different MOF materials are the main influencing factors of fluorescence quenching to recover efficiency. Finding better materials may be beneficial to further reduce the background signal and improve the sensitivity of the detection system. The use of luciferase as a reporter gene in the PC reporter system could be used for the rapid prototyping of arrays and microsystems in the future, but for field applications, electrochemical signals, and colorimetric readings may be more suitable. Key features of Vigilant include short run times, compatibility with rapid extraction protocols, and isothermal amplification, which make it a practical method for detecting viruses and pathogens that can be used for the large-scale screening of COVID-19 cases, but the identification of nucleic acid sequences is dependent on PAM sequences. If any DNA sequence can be identified in a PAM-independent manner, the utility of this platform will be extended to any nucleic acid sequence, and the detection process of pre-amplification and detection step separation and how to better avoid cross-reaction also needs to be further considered. FLASH-NGS enable high levels of multiplexing (thousands of targets) and the highly multiplex detection of antimicrobial resistance genes directly in a patient′s sample, but if the patient is infected with multiple microbial properties, it is not possible to determine whether the acquired resistance genes originated in a particular species.
CRISPR-Cas12a (Cpf1) (Figure 1B) contains a predicted RuvC-like endonuclease domain, which can cleave dsDNA under the guidance of gRNA; this domain is also closely related to the corresponding nuclease domain of Cas9 [71]. Unlike Cas9, Cas12a recognizes a distal 5′-T-rich PAM and generates PAM distal dsDNA breaks with staggered 5′ and 3′ ends. Cas12a also can recognize complementary ssDNA sequences in a PAM-independent manner and cleave it. Different from Cas9, Cas12 has collateral strand cleavage activity [30,72], as the target DNA sequence is present. Cas12 will release powerful, indiscriminate single-stranded DNA (ssDNA) cleavage activity. Cas12 has been widely developed due to its trans-cleavage activity.
In simple terms, the Cas12a/gRNA effector is used to identify the target DNA sequence, forming a Cas12/gRNA/target DNA ternary complex and then non-specific trans-cleavage fluorescence quenching reporter probe ssDNA [40,73]; finally, detection results are obtained through changes in the fluorescence signal [31,74]. Like (DETECTR) [30] and HOLMES [31], Laevis family bacteria-derived LbCas12a is used for detection by the above principle, and an isothermal amplification step of the sample sequence is also introduced to improve the sensitivity. There is also an ITP–CRISPR assay [32] that uses an electric field gradient to purify nucleic acids and accelerate DNA and RNA hybridization to speed up CRISPR analysis [75,76]. Cas12/gRNA, target DNA, and reporter ssDNA were confocalized into a 100 μL system. Changes in fluorescence intensity were used to report the results. This method was successfully applied to the rapid detection of SARS-CoV-2 RNA, and the total detection time from the original sample to the result was reduced by about 30–40 min, demonstrating its excellent diagnostic ability. Cas12a-FDet [33] combines the reaction system in a sealed reaction tube, avoiding the risk of aerosol contamination during amplification and transfer. Like Cas12a-FDet, a one-pot toolbox with precision and ultra sensitivity (OCTOPUS) concentrated the recombinase-polymerase amplification (RPA) reagents [34], crRNA, and ssDNA-FQ reporter genes in a single tube, but Cas12a protein was stored on the tube cap. RPA was pre-amplified 15 min after sample collection, and then the reaction tube was rotated to make the Cas12a protein enter the system. After the target sequence was identified, CRISPR/Cas12a was activated, and the fluorescence reporter gene was cleaved to generate fluorescence for detection. Not only the fluorescence quenching reporter ssDNA probe is used to report the detection result, but some systems use other detection methods, such as Cas12a-UPTLFA, combined with LFA technology and up-converting phosphor nanoparticle (UCP) label-coupled detection methods [35]. The method used a FAM-biotin probe (UPT will be conjugated to biotin), labeled ssDNA as the reporter, and the results were obtained by scanning using a UPT biosensor. The cutoff value was set according to the T/C values of the negative controls to designate the positive or negative results. When the target sequence is present, the ssDNA is cleaved, and the UPT conjugate flows along the flow band, causing a change in the T/C value, effectively improving the detection sensitivity. There also have some CRISPR/Cas system-based detection systems based on electrochemical biosensor platforms [77], which have the advantages of fast signal reading, simple platform, and low transducer cost. Recombinase-aided amplification (RAA)-based E-CRISPR [36] (Figure 4A) uses methylene blue (MB) to modify the ssDNA reporter gene via the Au–S covalent bond and assemble it on the working electrode. First, the sample is amplified by recombinase-aided amplification (RAA). Then the non-specific ssDNA trans-cleavage activity of the CRISPR/Cas12a was activated in the presence of the target sequence. The reporter gene modified by MB was cleaved, and the results were reported by wave voltammetry (SWV) to detect before and after the microelectrochemical signal, to achieve the purpose of nucleic acid detection. The electrical impedance spectrum (EIS)–CRISPR [37] (Figure 4B) fixed ssDNA on the gold electrode, limiting electronic communication between the electrode and the solution. In the presence of the target DNA, CRISPR/Cas12a binds to the target DNA and triggers its trans-cleavage activity, cracking the ssDNA on the gold electrode, accelerating electron transfer between the electrode and the solution. Finally, combined with EIS technology, the presence of the target sequence is identified by detecting subtle changes in the current on the electrode surface [78]. CRISPR/Cas12b [30] has the same non-specific trans-cleavage capability as CRISPR/Cas12a. However, the CRISPR/Cas12b system showed different target preferences during the trans-cleavage process, and the non-specific ssDNA trans-cleavage rate was higher when dsDNA was used as the target than ssDNA. In addition, CRISPR/Cas12b targets ssDNA substrates by cleaving the ssDNA probe independently of the PAM, whereas targeting dsDNA requires the 5′-TTN-3′ PAM site [79]. The one-hour low-cost multipurpose highly efficient system (HOLMESv2) is a nucleic acid detection platform based on the CRISPR/Cas12b system [38]. Cas12b, as a thermophilic Cas protein, can bind to the loop-mediated isothermal amplification (LAMP) program [80]. The optimal PAM sequence (5′-TTC-3′ and 5′-TAC-3′) and the optimal reaction temperature (target ssDNA 35–65 °C target dsDNA 45~55 °C) were discussed, and a one-step detection system was constructed that could be used to distinguish SNPs and accurately quantify the methylation degree of the target DNA. The Cas12b-based DNA detection (CDetection) platform also utilizes the trans-cleavage activity of Cas12b [39], which can simultaneously distinguish between HPV16 and HPV18, with a detection limit of 10nM without pre-amplification.
These detection platforms based on the CRISPR/Cas12 system mainly rely on their collateral strand cleavage activity, combined with the application of fluorescence quenching reporter genes, so that the detection results can be better presented. However, the specific recognition of target sequences still requires the participation of 5′ AT-rich PAM, and the selection of target sequences is still limited. In addition, the CRISPR/Cas12 system has a high tolerance to the first eight nucleotide mismatches (<4 nt) near PAM, making the off-target effect more inevitable. HOLMES tested 10 kinds of Cas12a proteins and finally selected the Trichospiraceae bacteria ND2006 Cas12a (LbCas12a) to be used in related research, which laid a foundation for subsequent research. DETECTR can quickly and specifically detect HPV in human patient samples, thus providing a simple platform for point-of-care diagnosis based on nucleic acids, which can be extended to rapidly detect any DNA sequence of interest with high sensitivity and specificity, but its practical application aspects have not been talked about. In ITP-CRISPR, an electric field gradient is used to control and influence the rapid CRISPR-Cas12 enzyme activity when the target nucleic acid is recognized. It takes about 30 min from the original sample extraction to obtain the result. However, due to the limitation of microflow chip design, the analysis of this detection system is currently limited to processing 10 μL original sample as input. This can affect the sensitivity of the system, and the sample cleavage and LAMP steps in the process need to be done manually, which is a hindrance to achieving high-throughput detection. The Cas12aFDet system has certain anti-interference ability in complex sample detection, and the use of constant room temperature as the reaction temperature and a single-tube detection process renders the system expected to be used for immediate detection and also explores a feasible efficient one-pot detection system. OCTOPUS is also a one-pot detection system, which omits the additional opening process to avoid practical inconvenience and possible cross-sample contamination. It can reach its detection limit of 1 CFU/mL in less than 50 min, but its cost is much higher than RT-PCR (7–8 times), making it difficult to use for widespread disease and epidemic control. A portable UPT biosensor is used in CAS12A-UPTLFA. Non-professionals can complete CAS12A-UPTLFA within 1.3 h as a POCT method, and it does not require expensive quantitative PCR instruments or microplate readers. It is versatile and can be prefabricated in batches. However, the successful detection rate at the lowest detection concentration is only 75%, which limits its practical application. RAA-based E-CRISPR converts target-recognition activity into detectable electrochemical signals, which greatly improves sensitivity. However, due to the deficiency of its detection limit, it is necessary to conduct pre-enrichment treatment before the RAA reaction in the detection of low-pollution samples. Whether it can be well combined with the amplification step still needs further exploration. EIS—CRISPR also converts target-recognition activity into a detectable electrochemical signal and can specifically identify different clinical isolates, but the main limitation of this system is its detection response time, whether LAMP can be combined with it, and shortening the time required for detection still needs to be further explored.
Cas13a (C2c2) (Figure 1C) belongs to the type VI CRISPR/Cas system [81], contains two nucleotide sequence binding domains (HEPN), and has single-stranded RNA (ssRNA) cleavage activity [82]. In 2016, it was discovered to have additional cleavage activity triggered by target RNA, making it widely used in nucleic acid detection [83].
CRISPR/Cas13-based biosensing systems include specific and high sensitivity enzymatic reporter gene unlocking (SHERLOCK) [40] and combinatorial arrayed reactions for multiplexed evaluation of nucleic acids (CARMEN) [42]. CARMEN integrated CRISPR/Cas13, reporter gene mixtures, and pre-amplified samples into the microarrays, using fluorescence microscopy to monitor the reaction of each microwell, enabling a high degree of multiple nucleic acid detection. This method can also be used to detect low-frequency cancer mutations in cell-free (cf) DNA fragments, as well as homozygous and heterozygous genotypes based on single-base differences. More importantly, this method has great potential for the large-scale diagnosis of SARS-CoV-2 [42]. An easy-readout and sensitive-enhanced (ERASE) band based on CRISPR/Cas13 is an improved version of the more sensitive lateral flow strip detection technology [43]. Optimizing the number of reporter RNA molecules, streptavidin, and secondary antibodies in the ERASE band, the cleaved reporter molecule could only be detected in the antibody capture band (C-band), and the biotin capture band (T-band) could not detect a cleaved reporter molecule, while the uncleaved reporter molecule could be detected on both bands. The improved lateral flow strip not only has higher sensitivity and specificity but also has more intuitive results. The disappearance of the T-band was regarded as the positive threshold. ERASE has great potential and needs to be further developed. Some researchers try to use the change in turbidity as the result of qualitative diagnoses, such as the liquid–liquid separation phenomenon (LLPS) (the critical value of polymer length exists in the polymer solution; when the polymer length reaches the critical value, the LLPS phenomenon will appear, the solution becomes turbid; when the polymer length is lower than the critical value, the solution components are evenly mixed, and the solution is clear) [84]. Combined with the collateral cleavage activity of Cas12a or Cas13a, the designed LLPS-Cas12a/Cas13a [44] (Figure 5A),cleaves long-chain nucleotides when the target sequence appears, renders the solution clear, and realizes detection through the change of solution turbidity. Besides the detection of pre-processed nucleic acid sequences, researchers have designed several detection systems that are not limited to pre-processed samples. For example, the light-up RNA aptamer signaling-CRISPR-Cas13a assay [45] (Figure 5B) enables the direct detect of pathogens based on the CRISPR/Cas13 system. This technology introduced a luminescent RNA aptamer in the Broccoli/DFHBI-1T complex [85]. After the aptamer was digested, the free DFHBI-1T only emits weak fluorescence; while the aptamers were combined with broccoli, the fluorescence of DFHBI-1Tis emit was more than 100 times than the free state [45]. Meanwhile, the bacteria RNA was identified by CRISPR/Cas13 to judge whether the pathogen was alive or not and quantifies the target RNA content to accurately quantify the live bacteria. This technology shows good sensitivity and specificity in the detection and quantification of live Bacillus cereus in food samples. Allosteric probe-initiated catalysis and the CRISPR-Cas13a system (APC-Cas) [41] (Figure 5C) requires only a small number of bacteria for detection. An allosteric probe (AP) containing three functional domains (aptamer domain, primer domain, and T7 promoter domain) was introduced [86,87]. The aptamer domain was used to identify specific strains; the primer domain was used to amplify the amplified sequence, and the T7 promoter domain was used to initiate RNA transcription and amplification. AP is inactivated in the absence of the target pathogen, but once the target pathogen is present, the aptamer domain specifically recognizes and binds to it, allowing the AP to expand from hairpin-like inactivated structures to active structures. With the participation of DNA polymerase and primer, dsDNA was generated using AP as a template strand, and then the target pathogen was replaced, and the first amplification was achieved. The secondary amplification was then performed with the participation of T7 RNA polymerase; the ssRNA was then recognized by the Cas13a/crRNA complex, which activated the trans-cutting ability to cleave a large number of surrounding RNA gene reporter probes to achieve the third amplification. Finally, the results were identified in combination with the generated fluorescence signal. The researchers successfully tested the specificity of Salmonella enteritidis using this technique. These two systems do not require the pre-processing of samples, further reducing the detection platform requirements, and are more conducive to on-site detection. The area of technology that combines the CRISPR/Cas system and detects pathogens directly still has great potential.
CRISPR/Cas13 also has collateral strand cleavage activity, allowing non-specific cleavage around RNA after recognition. Using this direct RNA recognition activity, researchers developed assays that do not require sample pretreatment, which greatly improves the utility of the CRISPR/Cas system. Furthermore, the high-sensitivity recognition of the target RNA sequence can better reduce the impact of the off-target effect. In the SHERLOCK system, reagents can be lyophilized for long-term storage and reconstituted on paper for field applications at a cost as low as $0.61 per test. However, combining the Cas13a detection reaction with paper-spotting significantly reduced the absolute signal of the readout, and the detection limit was also reduced by a factor of 10. CARMEN enables large-scale CRISPR-based diagnostics, and CARMEN can overcome the challenge of sequence heterogeneity at target sites through crRNA multiplexing, and the costs can be as low as $0.05/test. In the broader context of pathogen detection, discovery, and evolution, CARMEN and NGS complement each other. The ERASE system provides a powerful visualization tool for CRISPR detection and fulfills the possibility of on-site COVID-19 diagnosis with minimal equipment, but the presence of shallow T-bands that do not completely disappear in the result report may lead to some weakly positive samples being judged as negative; the sensitivity is slightly lower than that of precision fluorescence detection equipment. The LLPS-CRISPR system provides a simple and inexpensive way to implement CRISPR-based molecular diagnostics and circumvent the introduction of chemical labels on DNA or RNA molecules, which is more environmentally friendly and less expensive than other methods, but its detection limit is low, and whether the sensitivity can be improved by several orders of magnitude by using an appropriate amplification reaction requires further experimental verification, and evaluating the test results with the naked eye or simple equipment is prone to false negative results. The light-up aptamer-based-Cas13a system uses a one-pot assay for live pathogen detection, capable of detecting target bacteria as low as 10 CFU and accurately quantifying live bacterial content from 0 to 100%; the integration of CRISPR-Cas 13a and light-up RNA aptamers helps to create a reverse-transcription-free, nucleic-acid-amplification-free, and label-free method, further reducing equipment requirements, testing time, and cost. However, since the shutdown response of light-up RNA aptamers to target bacteria will greatly affect the detection of the system, finding more suitable light-up RNA aptamers needs further exploration. The APC-Cas system can easily detect Salmonella Enteritidis levels as low as 1 CFU, far exceeding the detection limit of 400 CFU of RT-PCR, and can effectively distinguish contaminated milk from pasteurized milk; it does not require bacterial isolation, nucleic acid extraction, or washing steps, and it costs as low as $0.86 per test. However, the system uses three enzymes in three steps, which makes the operation process complicated. Further simplifying the detection process without changing the detection sensitivity will greatly enhance its practical application value.
The Cas14 protein (Figure 1D) is a low-molecular-weight (400–700 amino acid) RNA-guided nuclease that recognizes target ssDNA and cleaves it without restriction sequences [88,89]. CRISPR/Cas14 also performs non-specific cleavages around ssDNA nuclease molecule like CRISPR/Cas12 [88].
Some nucleic-acid-detection platforms are based on CRISPR/Cas14, such as CMP [47] and TSPE-Cas14a [46] (Figure 6). In the CMP system, pre-amplified samples are firstly, and the CRISPR/Cas14 system-mediated activation of the unique collateral strand cleavage activity is used to cleave fluorescence quenching ssDNA reporter probes in the presence of the target sequence; then direct fluorescence reading is used for pathogens diagnosis. The tag-specific primer extension (TSPE)-based CRISPR/Cas14a pathogenic detection system designed a tag-specific primer that contains two domains (a primer sequence domain matching the target, and a tag sequence domain matching sgRNA). This assay does not require the purification of nucleic acid; the primer sequence domain amplifies and is enriched after matching the target sequence, and then the target nucleic acid is separated from the mixture by streptavidin-coated magnetic beads (the sequence-specific amplicons containing the tag sequence and biotin tag sequence are captured by streptavidin-coated magnetic beads). Identifying the tag sequence domain matching sgRNA activated the trans-cleavage activity of CRISPR/Cas14 and cleaved the fluorescence quenching reporter gene, resulting in an enhanced fluorescence signal and achieved detection. This diagnostic platform has a low detection limit (single cell or one aM), high sensitivity, and wide adaptability. CRISPR/Cas14 system-mediated non-specific collateral strand cleavage activity does not require a specific PAM sequence; CMP and TSPE-Cas14a can easily detect various DNA targets of pathogens by redesigning sgRNA.
Cas14a is established as the smallest class 2 CRISPR effector demonstrated to date for programmable RNA-guided DNA cleavage and can be detected without the restriction of PAM sequences. The recognition of ssDNA substrates by Cas14a requires a seed region that is different from the PAM region for complementation, and more than 2-nt mismatches will strongly inhibit the activity of Cas14a. This provides a huge possibility for its realization of high-fidelity detection, which greatly increases its practical application value. The CMP system provides a good model for the future accurate and direct detection of pathogens, especially single-stranded DNA viruses. Based on the collateral cleavage activity of CRISPR-Cas14a and TSPE, a general bacterial nucleic acid diagnostic platform has been developed, with a low detection limit (single cell or one aM), high accuracy (99%) and wide adaptability. However, this diagnostic platform is very time-consuming. In the future, combining isothermal amplification, lateral flow, and other technologies with Cas-TSPE to develop an easy-to-use integrated diagnostic method is waiting to be explored.
The type III CRISPR/Cas system is a multi-component and multi-pronged immune effector, and its reprogramming and use are more complicated due to its complex enzyme structure [90]. Relying on the unique internal signal-amplification mechanism of the type III CRISPR/Cas system [91], the type III CRISPR complex can be programmed to specifically recognize the virus RNA. After recognizing the target virus RNA, the cyclase domain produces about 1000 cyclic nucleotides (cOA); the cOA activate Csx1, cutting off the fluorophore connected to the quencher [23] (Figure 1E). Although these detection methods save time, the detection limits are relatively insufficient. When combined with pre-amplification, the time advantage will be lost to some extent, but it still proves that the type III system can also be used for rapid and sensitive detection.
Some nucleic acid detection systems based on the type III CRISPR are designed to detect viral RNA, such as MORIARTY [48] (Figure 7A) and SCOPE [50] (Figure 7B). MORIARTY was designed with a recombinant active Lactococcus lactis Csm (LlCsm) complex to synthesize cOA6 with the participation of divalent ions and ATP, and then cOA6 activated the RNase activity of Csm6. Then RNA-FAM was cleaved to cause changes in fluorescence intensity and finally achieve the detection. Compared with real-time PCR, it does not require expensive equipment and can even detect target RNA at 0.5 fM under non-amplification conditions. Combined with reverse transcription and RT-RPA pre-amplification steps, the detection limit can reach the aM level. SCOPE designed the TtCmr/crRNA complex targeting RNA by purifying the endogenous Cmr complex (TtCmr) from T. Thermophilus HB8. Upon the recognition of the target RNA, the complex generates cOA molecules, which then triggers the cleavage of the reporter RNA by TTHB144, resulting in a detectable fluorescence signal. In addition, the target RNA sequence required for TTHB144 activation is identical to the sequence required for cOA synthesis, allowing for the complete distinguishing of single-base differences. Combined with RT-LAMP and other pre-amplified sample methods, the detection limit can also reach the aM level. There is also a VmeCmr–NucC coupled assay that is based on the III-B CRISPR system from Vibrio metoecus, uses purified VmeCmr to activate NucC by cA3 generated during the activation of target RNA, and then cleaves the fluorescent reporter gene to report the detection results [49].
The nucleic acid detection system based on the type III CRISPR/Cas system is faster and can directly detect RNA samples. Combined with some pre-amplification steps, the detection limit can reach the aM level. However, due to its complex effector protein, its programming is more complicated, and it is rarely used in detection technology and needs further exploration. SCOPE is the first class 1-based CRISPR-Cas nucleic-acid-detection tool with high sensitivity and specificity, rapid detection, and flexibility. However, the detection system requires an additional amplification step, and the two-step method limits its application to high-throughput testing. Finding an efficient one-pot system can alleviate this situation. The MORIARTY assay also showed detection sensitivity consistent with RT-PCR and appeared to have a greater dynamic range than RT-PCR in detecting viral RNA, with a temperature range requirement of 37–42 °C for all steps, eliminating the need for expensive equipment and making it potentially compatible with low-cost hand-warmer-mediated heating solutions, but whether sample complexity reduces assay sensitivity and how can it be better combined with isothermal amplification steps to boost the signal-to-noise ratio are still waiting to be explored.
In recent years, there has been a global pandemic of COVID-19 [92,93]. Many researchers have also tried to use CRISPR/Cas systems to facilitate the detection of SARS-CoV-2 [94,95]. The CRISPR-based system not only has better detection sensitivity, but also has great advantages in various aspects such as detection threshold, shortening the detection time, improving the detection efficiency, and reducing the consumption of detection reagents [95,96].
Vigilant [27], based on CRISPR/Cas9, designed the sgRNA that targets the SARS-CoV-2 N gene. The detected LOD can reach 2.5 copies/μL, and the sensitivity and specificity are 96.4% and 100%, respectively. This method can achieve detection in 35 min and can be widely used in areas with relatively scarce resources, after pre-assembly with isothermal amplification.
The ITP-CRISPR assay is a detection system based on CRISPR/Cas12 [32]. In the rapid tests of SARS-CoV-2 RNA, the virus N and E genes are tested as target sequences (when N or E genes were detected, the test result was interpreted as positive), with sensitivity and specificity of 93.8% and 100%, respectively. This method uses less than 0.2 μL of reagents for detection and can achieve automatic nucleic acid extraction, accelerate, and enhance CRISPR enzymatic reaction for detection; the total detection time can be reduced to 30 min. Compared with real-time PCR, which needs 30 min to one hour to extract nucleic acid and which has a total detection time of nearly 3 h, the ITP-CRISPR assay has a huge advantage.
The CRISPR/Cas13 system is also be applied in SARS-CoV-2 detection. CARMEN [42] was a high-throughput detection system designed based on a microarray chip. In a test of 400 SARS-CoV-2 samples, its detection sensitivity could reach 99.7%. There is no doubt that its high-throughput detection capabilities can play a huge role in large infectious diseases such as SARS-CoV-2. The CRISPR/Cas13-based ERASE [43] is simple, fast, cheap, and convenient. Strip brightness was used as the identification standard, and the nucleoprotein (N) gene of SARS-CoV-2 was used as the target sequence. The sensitivity and specificity of the method were 90.67% and 99.21%, respectively, in 649 clinical samples. Although the visual measurement is slightly less sensitive, it is still very suitable for nucleic acid detection in a large population to facilitate the tracing of the source of infections [43].
The type III CRISPR/Cas system is also designed to detect SARS-CoV-2. For example, MORIARTY [48] selects the region segment within the spike (S) gene of the SARS-CoV-2 virus as the target RNA, while ensuring that the 3′-protospacer flanking sequence (3′-PFS) of the viral target RNA remained with the 5′-tag of LlCsm crRNA. The temperature range of all steps in the test is required to be 37 to 42 °C, which avoids expensive equipment and improves its on-site utilization value. SCOPE [50] designed a crRNA that targeted the E gene of SARS-CoV-2. Even target RNA at 1 nM can be detected through this system, and the fluorescent signal can be detected within a few seconds after incubation. After LAMP was pre-amplificated, sensitivity detection reached an aM level (10–18 M) in about 35 min (30 min pre-amplification + 5 min CRISPR reaction). The target RNA sequence required for activating TTHB144 needs to be highly consistent with that required for cOA production, thus reducing the false-positive rate. In addition, the temperature required for all detection steps is around 55 °C, which is also compatible with commercially available RNA polymerases. In a VmeCmr–NucC coupled assay, the detection limit can reach 8 fM when the SARS-CoV-2 N gene is used as the target sequence, and the determination is carried out at 37 °C, which has guiding significance for the field population screening of COVID-19.
The CRISPR/Cas system has attracted more and more attention since it was found to have specific recognition of target DNA or RNA sequences and/or non-specific collateral strand cleavage activity. CRISPR-based diagnostic systems work primarily by altering the gRNA to identify any desired target sequence. The combination of biosensing technology with CRISPR not only improves the sensitivity and specificity of existing detection technologies but also greatly reduces the time and cost, shortening the pathogen detection process. The diagnostic system is simple in sample processing, easy to operate, not limited by various environments, does not need special instruments or expensive reagents, and is expected to establish a wide range of real-time diagnosis platforms [97]. The detection capabilities of CRISPR also have certain limitations that have hampered its development. Firstly, detection based on sequence can not avoid interference from the off-target effect. For example, when using CRISPR/Cas9/sgRNA to identify the target sequence, mismatches in the proximal region of PAM were highly tolerated [98], and single mismatches in the interval region were fully tolerated by LshC2c2 [83]. Of course, there are also high-fidelity CRISPR/Cas systems developed for nucleic acid detection, such as HypaCas9 developed in 2017, which greatly improves the ability of targeted detection and reduces the off-the-target effect [99]. Future research should also develop in the direction of high-fidelity detection. Secondly, CRISPR-based nucleic acid detections are mostly qualitative results, unable to determine the pathogenic nucleic acid load of patients, and the pathogenic nucleic acid load is often closely related to the patient′s disease development. Further understanding of the load of the target pathogen can help predict the development stage of the disease and provide guidance for subsequent treatment. How to better combine quantitative detection with CRISPR is also the focus of future research. In addition, there are also problems of sample cross and aerosol contamination in the detection process. Measures such as a centralized detection reaction system in a tube or the pre-storage of the Cas protein in the tube cap can certainly reduce cross-contamination to a certain extent, but these measures are difficult to achieve in high-throughput detection systems [42]. Combining high-throughput testing with less cross-contamination is another challenge. In summary, CRISPR-based biosensing technology is a major innovation in detection technology that has already had an impact on detection and diagnostic capabilities in many areas and will have an even greater impact in the future. In the face of current and future infectious-disease outbreaks, CRISPR-based biosensing could dramatically improve our ability to diagnose and perform mass screening in populations. The prevalence of infectious diseases in modern society makes the living environment of human beings more severe. It is more likely that we will see more emerging infectious diseases like COVID-19 in the future; these will be widespread, highly prevalent, and harmful [100]. In the face of such a severe test, CRISPR-based detection technology will be a powerful weapon. In the natural environment of population susceptibility screening, more and more infectious diseases are developing in the direction of pathogenic higher evolution, resulting in rapid changes in the nucleic acid sequences of viruses and bacteria. In the face of the evolution of infectious diseases, nucleic-acid-sequence-detection technology based on the CRISPR/Cas system will become the key to preventing pandemics in the future. | true | true | true |
|
PMC9601109 | Hui-Yu Luo,Gan Li,Yu-Guo Liu,Yuan-Hao Wei,Jun-Bin Chen,Xiang-Fu Gu,Jia-Qi Tang,Yue Zhao,Chu-Hong Su,Ling-Yu Xiao,Fei Xiong,Zhong-Daixi Zheng,Shi-Ying Wang,Long-Ying Zha | The Accelerated Progression of Atherosclerosis Correlates with Decreased miR-33a and miR-21 and Increased miR-122 and miR-3064-5p in Circulation and the Liver of ApoE-/- Mice with Streptozocin (STZ)-Induced Type 2 Diabetes | 13-10-2022 | Atherosclerosis,type 2 diabetes (T2D),microRNA,cholesterol efflux | Atherosclerosis is a major risk factor for type 2 diabetes (T2D) mortality. We aim to investigate the changes in miR-21, miR-122, miR-33a and miR-3064-5p in circulation and the liver of ApoE-/- mice with streptozocin (STZ)-induced T2D. Twenty 5-week-old male ApoE-/- mice were randomly assigned to the control (n = 10) and T2D group (n = 10) and intraperitoneally injected with a citrate buffer and streptozotocin (STZ) (40 mg/kg BW) once a day for three consecutive days. The successfully STZ-induced T2D mice (n = 5) and control mice (n = 5) were then fed with a high-fat diet (HFD) for 34 weeks. Compared to the control mice, ApoE-/- mice with STZ-induced T2D had slower (p < 0.05) growth, increased (p < 0.05) total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C), decreased (p < 0.05) high-density lipoprotein cholesterol (HDL-C) in serum, reduced (p < 0.05) TC and sterol regulatory element-binding protein-2 (Srebp-2), elevated (p < 0.05) ATP-binding-cassette-transporter-A1 (Abca1) in the liver, aggravated (p < 0.05) atherosclerotic lesions in the aorta, downregulated (p < 0.05) miR-21 and miR-33a, and upregulated (p < 0.05) miR-122 and miR-3064-5p in serum and the liver. In addition, the aortic lesions showed a positive correlation with miR-122 (r = 1.000, p = 0.001) and a negative correlation with miR-21 (r = −1.000, p = 0.001) in ApoE-/- mice with T2D. In conclusion, T2D-accelerated atherosclerosis correlates with a reduction in miR-21 and miR-33a and an elevation in miR-122 and miR-3064-5p in circulation and the liver of ApoE-/- mice. | The Accelerated Progression of Atherosclerosis Correlates with Decreased miR-33a and miR-21 and Increased miR-122 and miR-3064-5p in Circulation and the Liver of ApoE-/- Mice with Streptozocin (STZ)-Induced Type 2 Diabetes
Atherosclerosis is a major risk factor for type 2 diabetes (T2D) mortality. We aim to investigate the changes in miR-21, miR-122, miR-33a and miR-3064-5p in circulation and the liver of ApoE-/- mice with streptozocin (STZ)-induced T2D. Twenty 5-week-old male ApoE-/- mice were randomly assigned to the control (n = 10) and T2D group (n = 10) and intraperitoneally injected with a citrate buffer and streptozotocin (STZ) (40 mg/kg BW) once a day for three consecutive days. The successfully STZ-induced T2D mice (n = 5) and control mice (n = 5) were then fed with a high-fat diet (HFD) for 34 weeks. Compared to the control mice, ApoE-/- mice with STZ-induced T2D had slower (p < 0.05) growth, increased (p < 0.05) total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C), decreased (p < 0.05) high-density lipoprotein cholesterol (HDL-C) in serum, reduced (p < 0.05) TC and sterol regulatory element-binding protein-2 (Srebp-2), elevated (p < 0.05) ATP-binding-cassette-transporter-A1 (Abca1) in the liver, aggravated (p < 0.05) atherosclerotic lesions in the aorta, downregulated (p < 0.05) miR-21 and miR-33a, and upregulated (p < 0.05) miR-122 and miR-3064-5p in serum and the liver. In addition, the aortic lesions showed a positive correlation with miR-122 (r = 1.000, p = 0.001) and a negative correlation with miR-21 (r = −1.000, p = 0.001) in ApoE-/- mice with T2D. In conclusion, T2D-accelerated atherosclerosis correlates with a reduction in miR-21 and miR-33a and an elevation in miR-122 and miR-3064-5p in circulation and the liver of ApoE-/- mice.
Atherosclerosis is a widespread chronic disease and a major risk factor for myocardial infarction, stroke, and ischemic gangrene, and is a leading cause of morbidity and mortality worldwide [1]. It is caused by both genetic and environmental factors with complex pathogenesis [2]. Studies have shown that atherosclerosis and diabetes mellitus (DM) are connected. DM is a group of carbohydrate metabolism disorders and is mainly featured by chronic hyperglycemia due to the defects of insulin secretion and/or insulin action. The International Diabetes Federation (IDF) estimates that 415 million people worldwide have diabetes, 91% of whom have type 2 diabetes (T2D) [3]. The increased risk and accelerated progression of atherosclerosis have been found in diabetic patients. For instance, adolescents and children with type I diabetes mellitus (T1D) exhibit the early development of atherosclerosis [4]. In type 2 diabetes (T2D) patients, coronary atherosclerosis is often accelerated with the enlargement of the necrotic core size, more aggravated inflammatory infiltrates, and more diffuse plaques in the coronary arteries [2]. In recent years, although the incidence of cardiovascular and cerebrovascular disease in diabetic patients has been decreasing, it is still the leading cause of death and disability in patients with T2D [5]. According to recent studies, 29.1% of patients with T2D had atherosclerosis [6]. Numerous studies have investigated the common pathological connections between atherosclerosis and diabetes mellitus [2,7]. Several factors including dyslipidemia with increased levels of atherogenic low-density lipoprotein (LDL), hyperglycemia, insulin resistance, oxidative stress and inflammation are proposed to explain the acceleration of atherosclerosis in diabetes mellitus. However, the pathological factors affecting the accelerated development of atherosclerosis in diabetes are still not completely understood. Understanding the underlying mechanisms is crucial for identifying new potential molecular targets [7]. In recent years, accumulating evidence has suggested the role of microribonucleic acids (miRNAs) in mediating the metabolic transition between atherosclerosis and diabetes. MiRNAs are endogenous, approximately 22 nucleotide-long non-coding RNAs that post-transcriptionally inhibit gene expression by altering the stability of messenger RNAs (mRNAs) and/or repressing mRNAs translation via pairing to the 3′-untranslated regions (3′-UTR) of target mRNAs of protein-coding genes [8]. MiRNAs have evolutionary importance in the modulation of gene expression since many of them are highly conserved across species. A single mRNA may be regulated by several various miRNAs, so each miRNA may have multiple mRNA targets [9]. MiRNAs are found in distinct biofluids like blood, saliva, and urine and show remarkable stability due to their packaging into vesicles [10]. The MiRNA pattern in biofluids will change under pathological conditions. The relationship between miRNAs and diseases has high complexity [11]. Thus, it is more promising to detect the combinations of multiple miRNAs (miRNA signature) rather than a single miRNA when discovering the integral role of miRNAs in diseases. MiRNAs have shown promise for use in diagnostic or prognostic tests in diseases such as atherosclerosis, T2D, etc. [12]. Atherosclerosis is mainly characterized by lipid accumulation and chronic inflammation in the arterial wall. Several miRNAs (miR-21, miR-122, miR-33a and miR-3064-5p) have been shown to be associated with lipid metabolism and inflammation, thus possibly participating in the development of atherosclerosis [9,10,11,12,13]. However, the role of these miRNAs in the development of atherosclerosis in the case of T2D is still far from being completely understood. Therefore, the objective of this study is to investigate the changing pattern of miR-21, miR-122, miR-33a and miR-3064-5p in the circulation and liver of ApoE-/- mice with streptozocin (STZ)-induced T2D and to identify their potential as biomarkers for screening cardiovascular risks in T2D.
All animal care and intervention procedures in this study were approved by the Southern Medical University Animal Care and Use Committee (no. SMUA2017003). Twenty 5-week-old male ApoE gene knockout (ApoE-/-) C57BL/6J (B6) mice (no. 312024300001195) were obtained from Shanghai Southern Model Biological Research Center (Shanghai, China). All mice were housed in standard cages in a room maintained at 22 ± 1 °C on a 12 h/12 h light/dark cycle and fed with a high-fat diet (HFD, no. D12451, Guangdong Medical Laboratory Animal Center, Guangzhou, China) containing 45% fat with an energy level of 4.73 kcal/g. Diets and sterile water were provided ad libitum. After 1 week of adaptive feeding, all mice were randomly assigned to the control (n = 10) and type 2 diabetes (T2D) group (n = 10). Following an intraperitoneal glucose tolerance test (IPGTT), mice in the T2D group were given an intraperitoneal injection of STZ (dissolved in 0.1 M citrate buffer, pH 4.5; catalog number. S0130, Sigma, St. Louis, MO, USA) once a day at a dose of 40 mg/kg body weight (BW) for three consecutive days. Meanwhile, mice in the control group received an intraperitoneal injection of 3 mL/kg BW citrate buffer (0.1 M, pH 4.5). After the three days of injections, the random blood glucose levels of all mice were detected once a day for another three consecutive days using a glucometer. The IPGTT was performed once again. Mice in the T2D group were considered diabetic by using the standards as follows: (1) the random blood glucose level ≥ 11.1 mmol/L, and (2) the area under the curve (AUC) of IPGTT > the average AUC of IPGTT of all mice in the control group. Based on these standards, five mice (n = 5) were finally judged as diabetic and the other five mice were excluded as non-diabetic. Correspondingly, only five mice (n = 5) from the control group were randomly selected and remained. All mice (n = 5 in the control group and n = 5 in the T2D group) were then fed with an HFD for 34 weeks. BW and feed consumption were recorded once a week. The random blood glucose was determined once a week. At the end of the feeding experiment, the mice were fasted overnight, placed in a closed box with a concentration control valve, anesthetized with ether, and blood samples collected, and then sacrificed for the collection of the other samples.
Three days before and after the intraperitoneal injection of STZ or citrate buffer experiment, mice were fasted overnight and intraperitoneally injected with 2 mg/g BW of glucose [14]. Blood glucose levels at 0, 15, 30, 60, 90 and 120 min after intraperitoneal glucose injection were measured using the Roche ACCU-CHEK Performa glucometer (Indianapolis, IN, USA). The blood glucose curve was plotted and the AUC was calculated using GraphPad Prism 5.0 software.
After the heart was perfused with phosphate-buffered solution (PBS), the entire aorta was exposed under the dissecting microscope, carefully separated in its entirety, and fixed with 4% paraformaldehyde for 48–72 h. After fixation, the adipose tissues around the aorta were fully peeled off. Then the aorta was opened longitudinally with ophthalmic scissors and then fixed to a black dish in a horizontally tiled manner by steel needles. The aorta was cleaned three times with PBS to remove any floating impurities. The aorta was stained with oil red O at room temperature for one hour, then rinsed with PBS three times and photographed with a microscope with a Canon camera [15]. The Image-Pro Plus 6.0 software was used to calculate the proportion of red lipid plaques in the entire aorta.
Following overnight fasting and anesthetization of the mice, blood was collected from the retro-orbital plexus, separated, and the serum stored at −80 °C. Commercially available enzyme-linked immunosorbent assay (ELISA) kits (ExCell Biotech, Shanghai, China) were used to determine the level of insulin in the serum. The levels of triglyceride (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C), in the serum and liver, were determined using the corresponding commercial kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) as per the manufacturer’s recommended protocols. The blood glucose level without fasting was randomly determined using the glucometer once a week.
Total RNA was extracted from samples using the TRIZOL reagent (Invitrogen, Carlsbad, CA, USA) by following the manufacturer’s instructions. Aliquots of 0.5 μg RNA were reverse transcribed to cDNA using the PrimeScript RT reagent kit (AG, Guangzhou, China). A stem-loop primer was used for microRNA reverse transcription, and a universal primer in the PrimeScript RT reagent kit was used for the reverse transcription of U6. Quantitative PCR (QPCR) was performed using the LightCycler 96 (Roche, Indianapolis, IN, USA). QPCR was performed using the SYBR Green Realtime PCR Master Mix (AG, Guangzhou, China) [16], and the reaction mixtures were incubated at 95 °C for 1 min, followed by 40 cycles of 95 °C for 2 s and 60 °C for 20 s and 70 °C for 10 s. Experiments were replicated at least three times. The primers used in this study were listed in Table 1. The relative expression of miRNAs was evaluated using the 2−ΔΔ Ct method and normalized to the expression of U6, respectively.
50 mg of liver tissue was taken from each mouse and placed in a 1.5 mL EP tube. The tissue was soaked in a liquid containing 1% phenylmethylsulfonyl fluoride (PMSF) and cut into pieces of about 3 mm × 3 mm using ophthalmic scissors. PBS supernatant was removed by centrifugation after shaking the EP tube, and this step was repeated three times. Then, 400 μL cold radioimmunoprecipitation assay (RIPA, KeyGEN Biotech, Nanjing, China) containing 1% PMSF and 1% phosphatase inhibitor was added to the tube and three steel balls were added. Then, the tube was put into a low-temperature grinding machine (Xavier Biotechnology, Wuhan, China), at 70 Hz, for 60 s, and run twice. The supernatant was collected after being centrifuged at 10,000 rpm for 5 min at 4 °C, and the protein concentration was determined by the bicinchoninic acid (BCA) method. Western blotting was performed as described earlier [17]. In simple terms, a 1.0% thick tris–glycine gel was prepared and liver tissue protein samples were added at a loading of 40 μg, and the marker was added. Electrophoresis conditions were 80–120 V. The proteins of the gel were then transferred to the nitrocellulose membrane (Bio-Rad) at a voltage of 100 V. The membrane was subsequently treated with 5% bovine serum albumin in TBST for 120 min at 25 °C to prevent non-specific reactions before incubating with an anti-Abca1 antibody (1:1000; catalog number. 96292s, Cell Signaling Technology, Danvers, MA, USA), or anti-Srebp-2 antibody (1:1000; catalog number. 28212-1-AP, Proteintech, Wuhan, China), or anti-β-actin antibody (1:5000; catalog number. 4967s, Cell Signaling Technology, Danvers, MA, USA) for 8 h at 4 °C. After further incubating with secondary antibodies (anti-rabbit IgG; 1:5000; catalog number. 7074S, Cell Signaling Technology, Danvers, MA, USA) for 120 min at 25 °C, the membrane was washed a final time. All protein bands were developed using enhanced chemiluminescence and visualized with a Tanon-5200 imaging system (Shanghai, China).
Statistical analysis was performed using the independent sample t-test or Pearson correlation analysis by using SPSS 20.0 statistical software (SPSS, Chicago, IL, USA). Results were presented as means ± standard deviation (S.D.). A p-value less than 0.05 (p < 0.05) was considered as significant.
The HFD-fed ApoE-/- mice are a well-established model for studying atherosclerosis. T2D usually accelerates the progression of atherosclerosis. In this study, we used STZ to induce T2D in the HFD-fed ApoE-/- mice in order to investigate the progression of atherosclerosis in T2D mice. Before STZ injection, the AUC of IPGTT was not significantly different (p > 0.05) between mice in the control and T2D groups indicating no difference in their glucose metabolism (Figure 1A,B). Following STZ injection for three consecutive days, five out of the ten mice in the T2D group had a significantly higher (p < 0.05) AUC of IPGTT than mice in the control group (Figure 1C,E,F) and had random blood glucose levels ≥ 11.1 mmol/L (Figure 1D). Therefore, these five mice were judged as T2D mice and were finally included in the T2D group. During the whole period (34 w) of the experiment, mice in the T2D group had significantly higher (p < 0.05) random blood glucose levels at all time points than mice in the control group (Figure 1H). At the end of the experiment, mice in the T2D group had significantly elevated (p < 0.05) fasting blood glucose levels (Figure 1G) and decreased (p < 0.05) fasting insulin levels (Figure 1I) than mice in the control group. Meanwhile, mice in the T2D group had a significantly higher (p < 0.05) HOMA-IR and a lower (p < 0.05) ISI compared to mice in the control group.
As seen in Figure 2A, the growth curve of mice increased gradually throughout the whole period of the experiment. Compared to the control mice, although the initial body weight (IBW) was not statistically different (Figure 2B), the STZ-induced T2D mice showed a slower growth speed with a statistically significant lower (p < 0.05) BW from week 24 to 34. Therefore, the STZ-induced T2D mice had a significantly lower (p < 0.05) final body weight (FBW) and body weight gain (BWG) compared to the control mice (Figure 2B). The feed consumption had no statistical analysis because the five mice in each group were housed in one cage and the food intake was recorded per cage (Figure 2C). Even so, it seems that the STZ-induced T2D mice consumed more feed than the control mice from weeks 24 to 34. These results indicated that the STZ-induced T2D mice had slower growth and reduced BWG compared with control mice.
Obvious atherosclerotic plaques often develop in the large arteries of ApoE-/-mice fed with an HFD. In this study, we analyzed the atherosclerotic plaques in the aorta and quantified the atherosclerotic lesions. As shown in Figure 3, the STZ-induced T2D mice had significantly more severe (p < 0.05) atherosclerotic lesions in the aorta compared to the control mice. These results suggest that the progression of atherosclerosis was accelerated in the ApoE-/- mice with STZ-induced T2D.
We measured the lipid profiles (TC, TG, LDL-C, and HDL-C) in both serum and liver tissues since it is well acknowledged that an HFD-induced disorder of lipid metabolism pivotally contributes to the progression of atherosclerosis. The ApoE-/- mice with STZ-induced T2D had significantly increased levels (p < 0.05) of TC and LDL-C and decreased levels (p < 0.05) of HDL-C in the serum (Figure 4C) compared to the control mice. Meanwhile, the ApoE-/- mice with STZ-induced T2D had significantly decreased levels (p < 0.05) of TC in the liver (Figure 4D). In addition, the levels of TG in the serum and liver were not significantly different (p > 0.05) between the T2D mice and the control mice (Figure 4C,D). These results suggest that the STZ-induced T2D ApoE-/- mice had increased TC and LDL-C levels in the blood but decreased cholesterol levels in the liver. Cholesterol homeostasis is to a large extent determined by the synthesis/uptake and efflux of cholesterol in the liver. Therefore, we further looked at the expression of key molecules that are responsible for regulating the synthesis/uptake and efflux of cholesterol. SREBPs are key transcriptional regulators of genes involved in cholesterol biosynthesis/uptake. As seen in Figure 4E,F, the protein expressive levels of precursor Srebp-2 (P- Srebp-2) and mature Srebp-2 (N- Srebp-2) were significantly reduced (p < 0.05) in the liver tissues of the STZ-induced T2D ApoE-/- mice compared to that of the control mice. Meanwhile, the protein expressive level of Abca1, which is an important regulator of HDL synthesis and responsible for reverse cholesterol transport, was significantly elevated (p < 0.05) in the liver tissues of the STZ-induced T2D ApoE-/- mice compared to that of the control mice. These results indicate that the cholesterol synthesis in the liver was decreased but its efflux was increased in the ApoE-/- mice with STZ-induced T2D.
Accumulating evidence shows that multiple circulating miRNAs are dysregulated in glucose and lipid metabolic disorders as well as their associated diseases such as atherosclerosis, etc. In this study, we analyzed several miRNAs in the blood and liver that might be involved in atherosclerosis. The expressions of miR-21 and miR-33a in the blood (Figure 4A) and liver (Figure 4B) were significantly downregulated (p < 0.05) in ApoE-/- mice with STZ-induced T2D. However, the expressions of miR-122 and miR-3064-5p in the blood (Figure 4A) and liver (Figure 4B) were significantly upregulated (p < 0.05) in ApoE-/- mice with STZ-induced T2D. Moreover, we analyzed the relationship between the circulating miRNA levels and aortic lesion areas using Pearson correlation analysis (Table 2). The aortic lesions were positively correlated with the circulating levels of miR-122 (r = 0.997, p = 0.001) and miR-3064-5p (r = 0.951, p = 0.004), and negatively correlated with the circulating levels of miR-21 (r = −0.857, p = 0.029) and miR-33a (r = −0.912, p = 0.011) among all mice. In addition, the aortic lesions showed a positive correlation with miR-122 (r = 1.000, p = 0.001) and a negative correlation with miR-21 (r = −1.000, p = 0.001) in ApoE-/- mice with T2D. Together, these results suggested that the aortic lesions correlate with the decreased miR-33a and miR-21 and increased miR-122 and miR-3064-5p in circulation of ApoE-/- mice with STZ-induced T2D.
Many studies showed that the progression of atherosclerosis is accelerated in the case of T2D [2,7,18]. In this study, ApoE-/- mice fed with an HFD developed distinct atherosclerotic lesions in the aorta suggesting that it is a good and reproducible model for studying atherosclerosis. STZ injection caused hyperglycemia and insulin resistance in the ApoE-/- mice indicating the successful induction of T2D. In ApoE-/- mice with STZ-induced T2D, the development of atherosclerotic plaques was aggravated, which concurs with previous reports [2,7]. The mechanisms underlying the T2D-associated acceleration of atherosclerosis are still not completely understood although several mechanisms have been proposed [2]. MiRNAs have shown great potential for diagnostic or prognostic tests in diseases including atherosclerosis. The differential miRNAs expression patterns between common atherosclerosis and atherosclerosis combined with T2D may serve as biomarkers for screening cardiovascular risks in T2D. Here, the data indicated that several miRNAs (miR-21, miR-33a, miR-122, and miR-3064-5p) were potentially associated with the development of atherosclerosis in T2D. Atherosclerosis is primarily characterized by lipid metabolism disorders and inflammation [19]. Several studies have investigated the role of miR-21, the most abundant miRNA in macrophages, in inflammation and lipid metabolism as well as in atherosclerosis. It is well-established that miR-21 played a key role in the inhibition of proinflammatory responses and the resolution of inflammation [20]. The absence of miR-21 results in accelerated atherosclerosis, plaque necrosis and vascular inflammation [21]. Therefore, miR-21 downregulation in circulation may favor vascular inflammation which creates suitable conditions for the formation of atherosclerotic plaques. Moreover, miR-21 reduced intracellular lipid accumulation induced by stearic acid in Hepa 1-6 cells by downregulating fatty acid-binding protein 7 (FABP7), a direct target of miR-21 [22]. Raitoharju et al. reported that the expression of miR-21 was significantly upregulated (fold changes 4.61) in atherosclerotic arteries compared to nonatherosclerotic left internal thoracic arteries [23]. Cengiz et al. found that the level of miR-21 in plasma was elevated in subclinical atherosclerosis in hypertensive patients compared to healthy controls [24]. However, Canfrán-Duque et al. reported that miR-21 depletion in macrophages triggered atherosclerosis plaque progression [25]. A more recent study from Telkoparan-Akillilar et al. found that the miR-21 level diminished 3.5 times in the blood samples of patients with atherosclerosis compared to healthy controls [26]. Results from animal models also observed downregulated miR-21 levels in microdissected fibrous caps of ruptured plaques [20]. Our present results indicated that miR-21 level in the blood and liver was downregulated in ApoE-/- mice with T2D. In STZ-induced T2D rats, miR-21 antagomir could improve insulin resistance and lipid metabolism disorder by upregulating the expression level of metalloproteinases 3 (Timp3) [27]. In diabetes-associated vascular dysfunction, miR-21 stimulated the proliferation of vascular smooth muscle cells (VSMC) by targeting specificity protein-1 (SPI) [28] and protected endothelial cells against high glucose-induced endothelial cytotoxicity probably by inhibiting the expression of the death domain-associated protein (DAXX) [29]. Although most studies supported that miR-21 has a negative association with lipid accumulation and inflammation and miR-21 downregulation favors the progression of atherosclerosis, there are still controversies regarding the role of miR-21 in atherosclerosis as well as in T2D-associated atherosclerosis. This could be due to the different populations, animal models, or sample sources used by different scientists resulting in the inconsistent observations of miR-21 levels. Furthermore, detailed investigations are awaiting to clarify the role of circulating miR-21 as a biomarker for atherosclerosis as well as T2D-accelerated atherosclerosis. MiR-122 is one of the most abundant miRNAs expressed in the liver of both mice and humans [30]. As a liver-specific miRNA, miR-122 plays not only a central role in liver development, differentiation, and homeostasis but also a crucial role in hepatic cholesterol homeostasis and fatty acid metabolism. The antagonism of miR-122 has been found to diminish the expression of several genes associated with lipid metabolism (acetyl-CoA carboxylase alpha, ACC1; acetyl-CoA carboxylase beta, ACC2, etc.) and cholesterol synthesis (sterol regulatory element-binding protein 2, SREBP2, etc.) in the liver [31]. Esau et al. found that miR-122 inhibition decreased plasma cholesterol levels resulting in the improvement of steatosis in HFD-fed mice [32]. Willeit et al. indicated that circulating miR-122 levels were increased in subjects with metabolic syndrome or T2D suggesting it is a marker for disorders of hepatic lipid metabolism [33]. In addition to cholesterol metabolism disorders, the pathophysiology of atherosclerosis is also characterized by chronic inflammation, oxidation, and apoptosis. The miR-122 antagonism represents a potential therapeutic approach for atherosclerosis since miR-122 promoted proinflammatory factors and oxidant injury in the liver and cardiovascular system [10]. The anti-apoptotic role of miR-122 inhibition has been evidenced by miR-122 inhibitor greatly suppressing the ox-LDL-induced apoptosis in human aortic endothelial cells. The expression of miR-122 was significantly elevated in aortic endothelial cells of HFD-fed ApoE-/- mice [34]. This agrees with our present results that the expression of miR-122 in circulation and the liver was also significantly increased in ApoE-/- mice with T2D compared to the control. MiR-122 elevation has been found to aggravate insulin resistance in hepatocytes by targeting insulin-like growth factor 1 receptor (Igf-1r), indicating that it has a possible contribution to the progression of atherosclerosis in T2D [35]. Moreover, Li et al. found that miR-122 primarily originated from circulation endothelial cells and monocytes, and its level was elevated in patients with acute myocardial infarction compared with patients with unstable angina [36]. Wang et al. reported that the circulating levels of miR-122 were significantly elevated in patients with severe coronary atherosclerosis. Moreover, the serum miR-122 levels were positively correlated with atherosclerotic severity [37]. A recent study investigated the relationship of several miRNAs including miR-122 with subclinical atherosclerosis in subjects with metabolic syndrome and found that miR-122 was positively associated with the cardio-ankle vascular index (CAVI) and correlated negatively with the aortic pulse wave velocity (AoPWV). In addition, age and triglycerides enhanced the prediction of the AoPWV by miR-122 [38]. Wu et al. reported that miR-122 was upregulated in the aortic intima and serum of ApoE-/- mice induced by an HFD, and miR-122 inhibition repressed the atherosclerotic plaque progression and vulnerable plaque formation in ApoE-/- mice [39]. Collectively, these studies support that miR-122 correlates positively with atherosclerosis and its severity and suggest that miR-122 is a good biomarker for predicting the development of atherosclerosis. However, there is one exception, where circulating miR-122 levels were significantly downregulated in coronary artery disease patients, according to a recent study by Mishra et al. [40]. Cholesterol homeostasis plays an important role in the progression of atherosclerosis. SREBPs are key transcription regulators of genes involved in cholesterol biosynthesis/uptake. The expression of Srebp-2 is insulin-dependent since it was observed to be decreased in insulin-deficient mice and increased with the increase in supplemental insulin dosage in rat liver cells [41]. In this study, STZ-induced T2D resulted in the decline of insulin in the serum of ApoE-/- mice. Thus, the hepatic levels of Srebp-2 (P- Srebp-2 and N- Srebp-2) were correspondingly decreased in ApoE-/- mice with T2D compared to the control mice. In addition, the elevation of miR-122 expression may contribute to the decrease in Srebp-2 because it has been shown that the high miR-122 expression could inhibit the Srebp-2 expression in the liver [42]. MiR-33a embeds within the introns of the Srebp2 gene and is co-transcribed when Srebp2 is activated. Therefore, the hepatic downregulated expression of Srebp-2 led to the decline of miR-33a as observed in the present study. MiR-33a has target genes involved in cholesterol export, such as Abca1, Abcg1 and Niemann-Pick C1 (Npc1) [43]. As a main miR-33a target gene, Abca1 is essential for the biogenesis of HDL in the liver and is responsible for the movement of free cholesterol out of the cell which is called reverse cholesterol transport (RCT). The miR-33a-deficient mice had significantly higher plasma HDL-C levels compared to wild-type C57BL/6 mice [44]. The inhibition of miR-33a increased hepatic Abca1 and circulating HDL by as much as 40% [45]. The anti-miR-33a therapy enhances RCT and regresses atherosclerosis in LDL-R knockout mice [44]. Therefore, miR-33a inhibition has been considered a therapeutic method for elevating HDL and exhibits potential in the prevention of atherosclerosis. In the present study, the levels of miR-33a in the blood and liver were both decreased in ApoE-/- mice with T2D compared to the control mice. Meanwhile, the increased hepatic Abca1 level along with the elevated serum levels of TC and LDL-C were observed in ApoE-/- mice with T2D. These results indicate that T2D inhibited the expression of miR-33a and increased the cholesterol efflux in the liver of ApoE-/- mice. Increased liver cholesterol efflux which led to the elevation of TC and LDL-C in circulation may contribute to the aggravation of atherosclerosis in T2D. The silencing and inhibition of miR-33a usually results in HDL elevation in circulation, according to existing studies [44,45]. Interestingly, the miR-33a in the liver and circulation was decreased while the HDL-C in circulation was also decreased in ApoE-/- mice with T2D. It has been shown that the HDL generated by miR-33a inhibition was functional, and the absolute levels of plasma HDL cannot truly reflect the functional HDL which is responsible for the removal of cholesterol from peripheral tissues into the feces for excretion. Anyway, this awaits further investigation regarding the changes of total HDL-C and its functional forms during the T2D-accelerated progression of atherosclerosis. Collectively, these results suggested that the downregulation of miR-33a promoted hepatic cholesterol efflux contributing to aggravated atherosclerosis in ApoE-/- mice with T2D. MiR-3064-5p is the mature isoform of miR-3064 located in chromosome 17q23.3 [13]. It has been shown that miR-3064-5p inhibits cementoblast differentiation [16] and regulates osteogenesis [46]. Furthermore, miR-3064-5p plays roles in the regulation of cancers including hepatocellular carcinoma [47], gastric cancer [13], bladder cancer [48], and breast cancer [49]. The expressions of miR-3064-5p were increased in both the blood and liver of ApoE-/- mice with T2D compared to the control mice. According to our unpublished study, miR-3064-5p targets the IκBα gene and activates NF-κB signaling in palmitate-stimulated RAW264.7 macrophage cell lines, indicating its role in the regulation of inflammation. Thus, miR-3064-5p possibly participates in the regulation of inflammation during the T2D-accelerated progression of atherosclerosis, which needs to be further elaborated.
In conclusion, STZ-induced T2D accelerates the progression of atherosclerosis in ApoE-/- mice. T2D-accelerated atherosclerosis correlates with decreases of miR-21 and miR-33a and the elevation of miR-122 and miR-3064-5p in circulation and the liver. This study adds a novel understanding of the potential roles of these miRNAs as biomarkers for predicting atherosclerotic progression in T2D. | true | true | true |
|
PMC9601115 | Elisa Orlandi,Elisa De Tomi,Rachele Campagnari,Francesca Belpinati,Monica Rodolfo,Elisabetta Vergani,Giovanni Malerba,Macarena Gomez-Lira,Marta Menegazzi,Maria Grazia Romanelli | Human Melanoma Cells Differentially Express RNASEL/RNase-L and miR-146a-5p under Sex Hormonal Stimulation | 11-10-2022 | mRNA/miR-146a interaction,testosterone,17β-estradiol,melanoma cells,gene regulation | Polymorphisms in the ribonuclease L (RNASEL) coding gene and hsa-miR-146a-5p (miR-146a) have been associated with melanoma in a sex-specific manner. We hypothesized that RNASEL and miR-146a expression could be influenced by sex hormones playing a role in the female advantages observed in melanoma incidence and survival. Thus, we explored the effects of testosterone and 17β-estradiol on RNASEL and miR-146a expression in LM-20 and A375 melanoma cell lines. Direct targeting of miR-146a to the 3′ untranslated region (3′UTR) of RNASEL was examined using a luciferase reporter system. Our results indicate that RNASEL is a direct target of miR-146a in both melanoma cell lines. Trough qPCR and western blot analyses, we explored the effect of miR-146a mimic transfection in the presence of each hormone either on RNASEL mRNA level or on protein expression of RNase-L, the enzyme codified by RNASEL gene. In the presence of testosterone or 17β-estradiol, miR-146a overexpression did not influence RNASEL transcript level in LM-20 cell line, but it slightly induced RNASEL mRNA level in A375 cells. Remarkably, miR-146a overexpression was able to repress the protein level of RNase-L in both LM-20 and A375 cells in the presence of each hormone, as well as to elicit high expression levels of the activated form of the extracellular signal-regulated kinases (ERK)1/2, hence confirming the pro-tumorigenic role of miR-146a overexpression in melanoma. Thereafter, we assessed if the administration of each hormone could affect the endogenous expression of RNASEL and miR-146a genes in LM-20 and A375 cell lines. Testosterone exerted no significant effect on RNASEL gene expression in both cell lines, while 17β-estradiol enhanced RNASEL transcript level at least in LM-20 melanoma cells. Conversely, miR-146a transcript augmented only in the presence of testosterone in either melanoma cell line. Importantly, each hormone acted quite the opposite regarding the RNase-L protein expression, i.e., testosterone significantly decreased RNase-L expression, whereas 17β-estradiol increased it. Overall, the data show that, in melanoma cells treated with 17β-estradiol, RNase-L expression increased likely by transcriptional induction of its gene. Testosterone, instead, decreased RNase-L expression in melanoma cell lines with a post-transcriptional mechanism in which miR-146a could play a role. In conclusion, the pro-tumor activity of androgen hormone in melanoma cells could be exacerbated by both miR-146a increase and RNase-L downregulation. These events may contribute to the worse outcome in male melanoma patients. | Human Melanoma Cells Differentially Express RNASEL/RNase-L and miR-146a-5p under Sex Hormonal Stimulation
Polymorphisms in the ribonuclease L (RNASEL) coding gene and hsa-miR-146a-5p (miR-146a) have been associated with melanoma in a sex-specific manner. We hypothesized that RNASEL and miR-146a expression could be influenced by sex hormones playing a role in the female advantages observed in melanoma incidence and survival. Thus, we explored the effects of testosterone and 17β-estradiol on RNASEL and miR-146a expression in LM-20 and A375 melanoma cell lines. Direct targeting of miR-146a to the 3′ untranslated region (3′UTR) of RNASEL was examined using a luciferase reporter system. Our results indicate that RNASEL is a direct target of miR-146a in both melanoma cell lines. Trough qPCR and western blot analyses, we explored the effect of miR-146a mimic transfection in the presence of each hormone either on RNASEL mRNA level or on protein expression of RNase-L, the enzyme codified by RNASEL gene. In the presence of testosterone or 17β-estradiol, miR-146a overexpression did not influence RNASEL transcript level in LM-20 cell line, but it slightly induced RNASEL mRNA level in A375 cells. Remarkably, miR-146a overexpression was able to repress the protein level of RNase-L in both LM-20 and A375 cells in the presence of each hormone, as well as to elicit high expression levels of the activated form of the extracellular signal-regulated kinases (ERK)1/2, hence confirming the pro-tumorigenic role of miR-146a overexpression in melanoma. Thereafter, we assessed if the administration of each hormone could affect the endogenous expression of RNASEL and miR-146a genes in LM-20 and A375 cell lines. Testosterone exerted no significant effect on RNASEL gene expression in both cell lines, while 17β-estradiol enhanced RNASEL transcript level at least in LM-20 melanoma cells. Conversely, miR-146a transcript augmented only in the presence of testosterone in either melanoma cell line. Importantly, each hormone acted quite the opposite regarding the RNase-L protein expression, i.e., testosterone significantly decreased RNase-L expression, whereas 17β-estradiol increased it. Overall, the data show that, in melanoma cells treated with 17β-estradiol, RNase-L expression increased likely by transcriptional induction of its gene. Testosterone, instead, decreased RNase-L expression in melanoma cell lines with a post-transcriptional mechanism in which miR-146a could play a role. In conclusion, the pro-tumor activity of androgen hormone in melanoma cells could be exacerbated by both miR-146a increase and RNase-L downregulation. These events may contribute to the worse outcome in male melanoma patients.
Melanoma incidence is continuously increasing in most developed countries [1]; epidemiologic data suggest that besides UV exposure, other factors such as sex, genetics, and epigenetics enhance the risk of developing cutaneous melanoma [2]. Sex differences in melanoma incidence and outcome have been constantly registered, yet not completely explained [3,4]. Male gender is indeed associated with a greater incidence of primary melanoma, and gender is an independent factor affecting survival in melanoma patients [5]. Complex molecular mechanisms and several exogenous and endogenous factors are involved in melanoma progression. One of which is the sex hormones milieu, which can influence gene expression by acting at both transcriptional and post-transcriptional levels [6,7,8]. Recently, melanoma has been reported to grow faster in male than in female mice [9]. Testosterone could likely be responsible for the worse melanoma prognosis in males. Indeed, it can induce melanoma cell growth via androgen receptor (AR) since genetic and pharmacological suppression of AR activity in melanoma cells blunts their proliferation, while increased AR expression or activation exerts opposite effects [10]. Remarkably, AR expression is positively correlated with poor survival of cutaneous melanoma patients [11]. AR, indeed, plays a key role in increasing melanoma cell invasion in multiple cell lines in vitro and in a mouse model in vivo [11]. Nevertheless, other authors claimed that testosterone supported melanoma cell growth also independently of AR binding, i.e., by inducing zinc influx and activating Mitogen Activating Protein Kinases (MAPK) [9]. At the same time, the melanoma protective effect in females could result from estrogen signaling through the G protein-coupled estrogen receptor (GPER) [12]. GPER activation in melanoma induces several phenotypic changes that inhibit tumor growth, and also render tumor cells more susceptible to clearance by native immune cells [12]. We previously observed a sex-specific interaction between rs486907 polymorphism in the ribonuclease L (RNASEL) gene and rs2910164 in miR-146a gene, showing that, only in the male population, allele rs2910164C represents a risk factor for individuals carrying the RNASEL rs486907GG genotype [13]. This sex-dependent linkage of both genes has been observed in non-melanoma skin cancer, as well [14]. Additionally, the same miR-146a polymorphism has been associated with various diseases in a sex-dependent fashion [15,16,17]. Micro RNA (miRNAs) are non-coding RNAs able to inhibit translation or induce mRNAs degradation by binding to their mRNA targets in specific complementary sequences usually located in the 3′ untranslated regions 3′(UTR), although they can bind also to other mRNA sequences, such as the 5′UTR or the coding region [18]. MiRNAs can be sex- and tissues-differentially expressed [19,20], suggesting their implication in sex bias in disease predisposition or outcome. Importantly, the different miRNAs expression level is a common attribute of malignant cell phenotype compared to normal one [21,22]. It should be noted that miR-146a is an important modulator of inflammation by affecting the innate and adaptive immune response [23]. Pan et al. reported that macrovesicles containing high level of miR-146a significantly promoted microvascular endothelial cell proliferation with the associated increase of the extracellular signal-regulated kinases (ERK)1/2 phosphorylation level [24]. In cancer, miR-146a can have either tumor promotive or tumor suppressive effects [25], although in melanoma it plays an oncogenic role [26]. RNASEL gene encodes the 2′,5′-oligoadenylate synthetase-dependent ribonuclease L (RNase-L), an enzyme that displays an antiviral role and may control the half-life of several mRNAs [27]. RNASEL/RNase-L regulate critical cellular functions including host antiviral response, apoptosis, and tumor-suppressive activity; thus, their expression must be tightly regulated [28,29,30]. The contribution of the RNASEL 3′UTR to the stability of its mRNA has been previously investigated [31]. Apart from several AU rich elements, RNASEL 3′UTR contains potential binding sites for some miRNAs, including a putative binding site for miR-146a [31]. In the present work, firstly we assessed whether RNASEL mRNA could be a direct target of miR-146a in LM-20 and A375 melanoma cells and in HaCaT cells. HaCaT is a human adult, low-calcium, high-temperature, keratinocyte cell line, which represents a cellular model commonly used to study normal keratinocytes [32,33,34]. The cells exhibit, similar to human keratinocytes, the following features: physiological keratinocyte morphology, epidermal differentiation capability, non-tumorigenic, and can undergo ultraviolet light-induced apoptosis [35]. Keratinocytes, indeed, can impact melanoma progression as they belong to tumor microenvironment by closely interacting with melanocytes in skin melanoma. Again, we investigated whether the transfection of miR-146a mimic in the presence of each sex hormone could affect the RNASEL/RNase-L expression level, as well as ERK1/2 activity (pERK1/2). Finally, we measured the endogenous levels of miR-146a and RNASEL/RNase-L expression in melanoma cells treated with either testosterone or 17β-estradiol. Our results suggest that sex hormones can influence RNase-L expression, i.e., testosterone decreased it by a post-transcriptional mechanism involving the increment of miR-146a gene expression, whereas 17β-estradiol was able to enhance RNase-L level by inducing RNASEL transcription.
Melanoma cell line LM-20 (17697M) derived from a male nodal metastasis [36] was provided by Dr Monica Rodolfo (Istituto Nazionale Tumori, Milan). A375 melanoma cells (CRL-1619; ATCC, Manassas, VA, USA) and LM-20 cells were grown in Roswell Park Memorial Institute 1640 (RPMI-1640, Gibco, BTL, Invitrogen Corp., Carlsbad, CA, USA). HaCaT cells were originally isolated from human adult skin and named based on their origin (human, adult, low-calcium, high-temperature, keratinocytes) [35]. HaCaT is considered a permanent non-malignant epithelial cell line, which maintains full epidermal differentiation capability [35]. Although it exhibits spontaneous phenotypic transformation and despite the unlimited growth potential, HaCaT cells, similar to normal keratinocytes, it reforms an orderly structured and differentiated epidermal tissue when transplanted onto nude mice, remaining non-tumorigenic [32]. Keratinocytes HaCaT cells (CLS, Cell Lines Service, Eppelheim, Germany) were grown in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, BTL, Invitrogen Corp. Carlsbad, CA, USA) containing 4.5 g/L D-glucose. All media were supplemented with 10% heat-inactivated fetal bovine serum, 1% L-glutamine (200 nM solution) and 2% Penicillin-Streptomycin (5000 I.U/mL and 5000 µg/mL solution, respectively) all purchased from Gibco, BTL, Invitrogen Corp. Carlsbad, CA, USA. All cell cultures were maintained in an incubator in a humidified atmosphere, of 5% CO2, at 37 °C.
A 777bp fragment of the RNASEL 3′UTR (NM_021133.4) was amplified using primers described in Table 1, which introduce an Xba I restriction site at the 5′ and 3′ of the amplified fragment. The amplified fragment was subcloned downstream of the firefly luciferase gene in the pGL3 promoter vector (3′UTR_RNASEL_WT). A 3′UTR_RNASEL_WT plasmid containing 5 nucleotide deletion in the consensus site for miR-146a was obtained by QuikChange Site-Directed Mutagenesis Kit (Stratagene, Agilent Technologies, La Jolla, CA, USA) (3′UTR_RNASEL_MUT). Primers used for amplification and mutagenesis are reported in Table 1. The plasmid sequences were confirmed through restriction enzyme mapping analysis and DNA sequencing (BMR Genomics, Padova, Italy). Transient transfection of the recombined vectors was performed using Lipofectamine 3000 (Invitrogen Corp., Carlsbad, CA, USA), according to the manufacturer’s instructions. The two melanoma cell lines and HaCaT keratinocytes (1.3 × 105 cells) were seeded in 24 well plates. After 24 h, 1 µg of 3′UTR_RNASEL_WT or 1 µg of 3′UTR_RNASEL_MUT constructs, together with 50 nM miR-146a mimic and 5 ng of Renilla luciferase plasmid (as normalizer) were transfected. The pGL3 vector without insert was used as a negative control. After 24 h, the cells were lysed with Passive Lysis Buffer (Promega Italia srl, Milan, Italy), and the relative luciferase activity was assessed with the Dual-Luciferase Assay Reporter System (Promega Italia srl, Milan, Italy) [37].
For qPCR, cells were plated at 3 × 105 cells per well in a 6 well plate in 2 mL of RPMI phenol red-free supplemented with 10% charcoal treated serum, 1% L-glutamine (200 nM solution) and 2% Penicillin-Streptomycin (5000 I.U/mL and 5000 µg/mL solution, respectively) for 48 h. For western blot, seven thousand five hundred (7.5 × 105) cells were seeded in T25 flasks and incubated in 5 mL of RPMI phenol red-free supplemented with 10% charcoal treated serum, and 1% L-glutamine (200 nM solution) and 2% Penicillin-Streptomycin (5000 I.U/mL and 5000 µg/mL solution, respectively). Next, testosterone or 17β-estradiol (Sigma Aldrich, Milan, Italy), at the final concentration of 10−6 M and 5 × 10−7 M, respectively, was added to the medium and cells were cultured for the following 24 h. Cells incubated with the corresponding amounts of pure ethanol (<0.0001%) served as a negative control and were set to 1 for analysis. Hormone concentrations were chosen based on experiments described by Kanda et al. [38] and after cell viability testing. Cells were subsequently lysed with Trizol Reagent (Thermo Fisher Scientific, Milano, Italy) for RNA extraction or RIPA buffer for western blot (see below).
Total RNA for gene expression analyses was extracted with Trizol Reagent (Thermo Fisher Scientific, Milan, Italy), following the manufacturer’s protocol. RNA quantification was performed with Nanodrop 2000 spectrophotometer (Thermo Fischer Scientific, Milan, Italy), and 500 ng of RNA were retrotranscribed by SensiFAST cDNA Synthesis Kit (Bioline, Trento, Italy), following the manufacturer’s protocol. For miRs expression analyses, 1 µL for each sample of the total RNA was direct reverse transcribed with the TaqMan Advanced miRNA cDNA Synthesis Kit (Thermo Fisher Scientific, Milan, Italy), following the manufacturer’s protocol. After reverse transcription, RNASEL and miR-146a expression levels were determined by real-time polymerase chain reaction (qPCR) as described [39]. Normalization was performed using the TATA box protein (TBP) which showed to be the most stable gene in our experimental conditions [40]. Primers used for amplification are described in Table 1. For miRNAs analysis, TaqMan Fast Advanced Master Mix and the specific probe for miR-146a (Thermo Fisher Scientific, Milan, Italy) were used. Normalization of expression was performed by miR-191, which showed to be the most stable miRNA in our experimental conditions [40]. The Real-Time PCR was performed in the Bio-Rad CFX Connect Real-Time System using the SensiFAST SYBR no-rox Kit (Bioline, Trento, Italy) or the TaqMan Advanced microRNA probes (Thermo Fisher Scientific, Milan, Italy). Relative quantification was calculated by Pfaffl’s formula [41]. Each measurement was carried out in triplicate in at least three different experiments.
After incubation with hormones, cells were washed twice with ice-cold Phosphate-buffer solution (PBS) and then collected by scraping. Subsequent protein extraction and western blotting were performed essentially as previously described [42]. Briefly, cells were lysed in RIPA buffer and a mixture of protease inhibitors (Mirus Bio LLC, Medison, WL, USA). Total protein concentration was determined by detergent compatible (DC) Bradford Assay analysis (Thermo Fisher Scientific, Milan, Italy). Cellular proteins were separated in SDS polyacrylamide 10% gel electrophoresis (SDS-PAGE) and transferred onto PVDF membrane (Thermo Fisher Scientific, Milan, Italy) by a wet electrophoretic transfer method. Membranes were incubated for 2 h at room temperature with primary antibodies diluted in blocking buffer: the anti-RNase-L (1:10,000, Wuhan Fine Biotech Co, Dublin, CA, USA), the anti-phospho-ERK1/2 (1:2000, Cell Signaling Technology, Danvers, CO, USA), and to normalize the protein amount loaded in the gel, the anti-β-Actin (1:10,000, Wuhan Fine Biotech Co., Dublin, CA, USA). Images were acquired with Azure C300 Processing machine (Azure Biosystem, Dublin, CA, USA). The intensity of bands was quantified with ImageJ software.
After hormones incubation, miR-146a mimic (Sigma Aldrich, Milan, Italy) was transfected at a concentration of 50 nM using Metafectene (Biointex, München, Germany), according to the manufacturer’s recommendations and incubated for 24 h. Cells incubated only with hormones served as negative controls and were set to 1 for analysis.
All the results are reported as a mean value ± standard deviation (S.D.). Differences were analyzed with GraphPad Prism statistical program, using unpaired, two-tailed Student’s t-test. One asterisk, * p < 0.05; two asterisks, ** p < 0.01. For each type of experiment, a minimum of three independent biological replicates were performed. Normal distribution of data was tested using the Shapiro–Wilk test.
To evaluate the silencing role of miR-146a by targeting the consensus sequence in the 3′UTR region of the RNASEL messenger, we performed a luciferase transfection assay. Recombinant plasmids including a fragment of 777bp of RNASEL 3′UTR region containing a putative site for miR-146a binding (wild type, WT), as well as its mutated sequence (MUT), were cloned. LM-20, A375 and HaCaT cell lines showed a significant down-regulation in the relative luciferase reporter activity when transfecting 3′UTR_RNASEL_WT construct in the presence of 50 nM miR-146a mimic. The corresponding mutated plasmid 3′UTR_RNASEL_MUT restored in part the activity of the luciferase (Figure 1). Thus, the results show that RNASEL mRNA is a target of miR-146a in the analyzed cell lines.
Firstly, as a transfection efficiency control, the level of miR-146a expression in cells transfected with miR-146a mimic has been measured. Figure 2A shows higher miR-146 amount in cells transfected with miR-146a mimic in comparison with each negative control in both melanoma cell lines. Overexpression of miR-146a in LM-20 cells did not affect the expression of RNASEL gene in the presence of each hormone. (Figure 2B, left). Conversely, miR-146a mimic transfection slightly increased RNASEL gene expression level in A375 cells in the presence of testosterone or 17β-estradiol (Figure 2B, right). As expected, miR-146a overexpression resulted in a lower RNase-L protein amount in both cell lines cultured in the presence of testosterone, as well as in the presence of 17β-estradiol (Figure 2C). To further demonstrate a pro-tumorigenic effect of miR-146a overexpression in melanoma cells, immunoblots were performed to measure the activated form of ERK1/2 enzymes. As shown in Figure 2D, miR-146a mimic transfection significantly increases the phosphorylated and active form of ERK1/2 in both A375 and LM-20 melanoma cell lines.
Once the functional role of miR-146a on RNASEL/RNase-L expression has been validated by transfection experiments, we investigated the effect of the hormonal milieu on RNASEL, miR-146a, and RNase-L expression directly in cells. Melanoma cells were incubated with testosterone or 17β-estradiol for 24 h and subsequently the expression levels of miR-146a, RNASEL and RNase-L were evaluated by qPCR or Western Blot analyses.
The transcriptional level of RNASEL gene was measured by qPCR. In both melanoma cells, the presence of testosterone showed no significant effect on RNASEL gene expression. In contrast, 17β-estradiol increased RNASEL mRNA level in LM-20, thus supporting a transcriptional induction of the gene (Figure 3, Top, left panel).
The incubation with 17β-estradiol did not modulate miR-146a gene expression in either LM-20 nor in A375 cell lines. Conversely, testosterone induced a significant upregulation of miR-146a transcript in both LM-20 and A375 melanoma cell lines (Figure 3, Top, right panel).
The effects of sex hormones on the two different human melanoma cells were investigated by analyzing the expression level of RNase-L protein by western blot. The expression of RNase-L was significantly diminished in both LM-20 and A375 melanoma cells (Figure 3, Bottom). On the contrary, 17β-estradiol administration increased the RNase-L protein amount in both melanoma cell lines (Figure 3, Bottom).
Gene expression analysis in different non-reproductive tissues shows the presence of a gender-specific effect on gene transcription [43], and on gene epigenetic regulation [44]. Accumulating evidence demonstrate the advantages of women in various types of cancer, including colorectal cancer, urothelial, and kidney cancer [45,46], as well as melanoma [47]. In melanoma, differences between males and females in tumor thickness, anatomic location, and level of invasion are some of the factors that have been linked to the progression of malignancy and survival of patients suggesting explanations for the female melanoma advantage [48,49,50,51]. Sex hormone signaling can affect cancer predisposition through several mechanisms, influencing tumor microenvironment, immune system, and the overall metabolic balance of an organism [8,9,10,11,48]. We had previously shown that polymorphisms of RNASEL and miR-146a genes were associated with melanoma in a sex-specific manner [13]. The present work intended to investigate, in sex-hormone-treated melanoma cell lines, the contribution of RNASEL, its product RNase-L enzyme, and miR-146a in the female advantages observed in melanoma incidence and survival. It is well known that RNASEL 3′UTR region displays a putative binding site for miR-146a [31]; however, the functionality of this site was not proven so far. At the present point, we attest that RNASEL mRNA is a target of miR-146a since the binding of miR-146a to RNASEL 3′UTR region lowers the luciferase reporter activity in two melanoma cell lines and non-transformed keratinocytes, as well (Figure 1). Moreover, overexpression of miR-146a by mimic transfection in the presence of either testosterone or 17β-estradiol can decrease RNase-L protein expression without interfering with RNASEL mRNA level in LM-20 melanoma cells (Figure 2). This suggests that miR-146a was able to decrease RNase-L protein amount likely by binding to RNASEL transcript and destabilizing its translation. Dissimilar results were obtained with miR-146a mimic transfection in A375 melanoma cells after treatment with each hormone, in which RNASEL transcript level was slightly upregulated (Figure 2B). This unexpected data could be due to an indirect downregulation of a miR-146a-target repressor factor acting on RNASEL gene promoter or by a direct activation performed by miR-146a itself on RNASEL gene transcription, as recently described [52]. In any case, this result needs further insights in the future. Nevertheless, at the protein level in the presence of each hormone, miR-146a mimic drives a slight but significant decrease in the amount of RNase-L enzyme, suggesting that the inhibitory effect of miR-146a could be dominant on RNASEL transcriptional activation (Figure 2C). In cancer, through different signaling pathways, miR-146a can affect the expression of several cancer-related genes and ultimately have dual conflicting roles in tumor progression and metastasis [25]. In fact, miR-146a displayed tumor suppressive functions in pancreatic cancer cells and colon cancer [53,54]. Conversely, in melanoma, miR-146a plays an oncogenic role. Indeed, it is upregulated by the oncogenic kinases BRAF and NRAS, and its overexpression in melanoma cells promotes proliferation and leads to tumor formation when overexpressing cells are injected in mice [26,55]. In concert, our data showed an increase in the phosphorylation level of the ERK1/2 MAPK in both melanoma cell lines transfected with miR-146a mimic (Figure 2C). Thus, miR-146a overexpression, in sex-hormones treated melanoma cells, supports the oncogenic role of miR-146a in melanoma by downregulating RNase-L expression and by activating the growth-promoting kinases ERK1/2. Subsequently, we investigated in the context of melanoma cells whether testosterone or 17β-estradiol administration may affect the endogenous expression of RNASEL and miR-146a genes and RNase-L protein. Our data attest that the expression of miR-146a was significantly increased by testosterone in two melanoma cell lines (Figure 3). The augmented expression of miR-146a may represent a disadvantage in males since it can act as a negative regulator of immune activation and can negatively interfere, among others, with the signal transduction and activator of transcription (STAT)-1/interferon (IFN)-γ axis, which in turn can reduce cell-migration, cell cycle activity, and basal oxygen consumption rate in melanoma cells [56,57]. Furthermore, miR-146a promoted primary melanoma progression by activating Notch signaling as described by Forloni et al. [26]. RNASEL gene expression in the presence of either hormone displayed only little difference compared to untreated control cells. A very slight but significant increase in RNASEL transcript was registered only with 17β-estradiol treatment in LM-20 cells (Figure 3). The majority of variations emerged when RNase-L protein expression was assessed. We observed in both melanoma cell lines an increased amount of RNase-L protein after cell incubation with 17β-estradiol and a lower enzyme expression with testosterone treatment (Figure 3). We can speculate that a lower RNase-L protein amount caused by testosterone could result from a post-transcriptional downregulation of the gene that could be driven, at least in part, by the concomitant upregulation of miR-146a. Whereas, the higher expression of the enzyme in 17β-estradiol-treated cells could be a consequence of transcriptional RNASEL induction. For its endo-ribonucleolytic activity, RNase-L regulates cell growth and differentiation by cleaving and destabilizing several mammalian RNA targets [58]. Notably, RNase-L targets include many miRNAs transcripts as well. Thereby, RNase-L can act as tumor-suppressor in mammalian cells also via destabilization of the miRNA-regulated transcriptome [58]. This multiplicity of targets, which is a typical feature of other antitumor RNases [59], results in pleiotropic effects, ranging from cell growth inhibition to pro-apoptotic activity [57], and suggests that RNase-L can play a protective effect on tumor progression and participate in the survival advantage in female melanoma patients. Remarkably, RNase-L negatively regulates androgen signaling by a protein-protein interaction mechanism [60]. Again, the high level of RNase-L expression in the presence of 17β-estradiol could result in further suppression of androgenic response; conversely, the low expression level in testosterone-treated cells can further potentiate AR signaling. Androgen signaling, indeed, has been demonstrated to promote cancer progression by inhibiting apoptosis or by promoting cell migration and matrix metalloproteinase activity in prostate cancer cells [60]. As previously reported [10], the major committed for the worse melanoma prognosis of males versus females could be the higher level of testosterone, which in turn activates AR. Genetic or pharmacological suppression of AR activity in melanoma cells hinders their proliferation [10]. Very recently, Vellano et al. demonstrated that AR blockage promoted a better response to BRAF/MEK-targeted therapy in melanoma patients, thus improving recurrence-free survival [61].
In conclusion, the results obtained in this study suggest that sex hormones may act on miR-146a, on RNASEL gene expression, and especially on RNase-L protein level with a view to contribute to the female advantages observed in melanoma incidence and survival. | true | true | true |
|
PMC9601231 | Adriana Solis-Vivanco,Mónica Santamaría-Olmedo,Dalila Rodríguez-Juárez,Margarita Valdés-Flores,Carlos González-Castor,Rafael Velázquez-Cruz,Eric Ramírez-Salazar,Ana Cristina García-Ulloa,Alberto Hidalgo-Bravo | miR-145, miR-92a and miR-375 Show Differential Expression in Serum from Patients with Diabetic Retinopathies | 21-09-2022 | diabetic retinopathies,proliferative diabetic retinopathy,diabetic macular edema,microRNAs | Diabetic retinopathies are important disabling conditions. Micro-RNAs (miRNAs) are regulators of gene expression and diseases can change their expression. Our aim was to analyze the expression of miRNAs in serum and vitreous samples from patients with diabetic retinopathies. The following groups and number of individuals were included: proliferative diabetic retinopathy (PDR) (n = 16), diabetic macular edema (DME) (n = 17), and idiopathic epiretinal membrane (IEM) as non-diabetic controls (n = 23). The initial miRNA expression was explored using TaqMan low-density arrays (TLDAs) with subsequent validation through a quantitative polymerase chain reaction (qPCR). Target genes were identified through bioinformatic tools for enrichment analysis. The TLDAs revealed the following miRNAs with differential expression in terms of PDR vs. IEM: miR-320a-3p, miR-92a-3p, and miR-375-3p in the serum, with miR-541-5p and miR-223-5p in the vitreous samples. DME vs IEM: miR-486-5p, miR-145-5p, miR-197-3p, and miR-125b-5p in the serum, and miR-212-3p in vitreous samples. PDR vs. DME: miR-486-5p, miR-100-5p, miR-328-3p, miR-660-5p, and miR-145 in the serum and none in the vitreous samples. Validation was confirmed only for miR-145, miR-92a, and miR-375 in the serum. The relevant enriched pathways for these three validated miRNAs, miR-145, miR-92a, and miR-375 were the vascular endothelial growth factor and its receptor, hepatocyte growth factor receptor, epidermal growth factor, focal adhesion, and phosphoinositide 3-kinase. Our results support the involvement of miRNAs in the pathophysiology of diabetic retinopathies and reinforce their potential as biomarkers or therapeutic resources. | miR-145, miR-92a and miR-375 Show Differential Expression in Serum from Patients with Diabetic Retinopathies
Diabetic retinopathies are important disabling conditions. Micro-RNAs (miRNAs) are regulators of gene expression and diseases can change their expression. Our aim was to analyze the expression of miRNAs in serum and vitreous samples from patients with diabetic retinopathies. The following groups and number of individuals were included: proliferative diabetic retinopathy (PDR) (n = 16), diabetic macular edema (DME) (n = 17), and idiopathic epiretinal membrane (IEM) as non-diabetic controls (n = 23). The initial miRNA expression was explored using TaqMan low-density arrays (TLDAs) with subsequent validation through a quantitative polymerase chain reaction (qPCR). Target genes were identified through bioinformatic tools for enrichment analysis. The TLDAs revealed the following miRNAs with differential expression in terms of PDR vs. IEM: miR-320a-3p, miR-92a-3p, and miR-375-3p in the serum, with miR-541-5p and miR-223-5p in the vitreous samples. DME vs IEM: miR-486-5p, miR-145-5p, miR-197-3p, and miR-125b-5p in the serum, and miR-212-3p in vitreous samples. PDR vs. DME: miR-486-5p, miR-100-5p, miR-328-3p, miR-660-5p, and miR-145 in the serum and none in the vitreous samples. Validation was confirmed only for miR-145, miR-92a, and miR-375 in the serum. The relevant enriched pathways for these three validated miRNAs, miR-145, miR-92a, and miR-375 were the vascular endothelial growth factor and its receptor, hepatocyte growth factor receptor, epidermal growth factor, focal adhesion, and phosphoinositide 3-kinase. Our results support the involvement of miRNAs in the pathophysiology of diabetic retinopathies and reinforce their potential as biomarkers or therapeutic resources.
Diabetes mellitus (DM) is one of the most prevalent diseases worldwide. In Mexico, visual disability occupies the second-highest place among diabetic complications, affecting 40% of patients [1]. Proliferative diabetic retinopathy (PDR) and diabetic macular edema (DME) represent the main ophthalmic complications. PDR progresses to retinal detachment and neovascular glaucoma, while DME implies central vision impairment. Currently, the diagnosis of PDR and DME is made under observation with a slit lamp. Despite the progression in treatment strategies, the pathophysiology of ophthalmologic complications related to DM is not fully understood. In recent years, microRNAs (miRNAs) have emerged as important regulators of gene expression in approximately 60% of human genes. miRNAs are non-coding RNA molecules of 21 to 24 nucleotides that have the ability to modulate the availability of a protein. They are involved in the posttranscriptional regulation of gene expression through mRNA degradation or translation repression [2,3]. miRNAs participate in most of the biological processes, both normal and pathological [4]. In addition, miRNAs can be detected in several bodily fluids, which makes them attractive potential biological markers [5,6,7]. In the last decade, some studies have identified miRNAs that are differentially expressed in patients with certain eye diseases [7,8,9]. The evidence has demonstrated the differential expression of miRNAs in the serum, vitreous, and other eye tissues in relation to a pathological condition affecting different structures in the eye [10]. Focusing on diabetic complications affecting the retina, some miRNAs have been found with differential expression in the vitreous and serum samples [11]. Qing et al. proposed a special serological sieve of microRNAs that can serve as a “signature” of PDR, which could precede early diagnosis [12]. The identification of these miRNAs might provide additional knowledge regarding the genes and metabolic pathways undergoing modification during the progression of the disease. The ultimate aim is the discovery of potential disease markers and therapeutic targets. Their dysregulation in metabolic diseases underlines their potential as therapeutic targets [13]. At the same time, they can help us to understand the disparity in severity and the development of complications between individuals. Identifying the miRNAs with differential expression in diabetic patients would represent a link in the chain of knowledge that is required for the creation of new therapeutic strategies. Cumulative evidence supports the theory that miRNAs could serve as biomarkers for supporting diagnosis and follow-up. The aim of this study was to analyze the expression of miRNAs in the serum and vitreous samples from diabetic patients with established retinal microangiopathy.
Patients attending the ophthalmology department at the National Institute of Rehabilitation in Mexico City from 2018 to 2019 were invited to participate. We recruited 16 type-2 diabetic patients with PDR who showed high-risk characteristics without DME, 17 type-2 diabetic patients with severe diffuse DME, and 23 non-diabetic patients with idiopathic epiretinal membrane (IEM). Individuals with IEM were chosen as controls because it is a pathology not associated with a metabolic disorder and that requires vitrectomy as treatment since we cannot obtain vitreous samples from healthy patients without justification. All participants were older than 35 years old, with a recommendation for pars plana vitrectomy. All individuals were of Mexican origin, with at least two generations of Mexican ancestors. Exclusion criteria included those patients who had had anti-VEGF treatment six months before. Patients who had vitreous hemorrhage at the time of surgery were also excluded, to avoid contamination of the vitreous content of miRNAs. In the case of the RDP group, we were especially careful in selecting patients who required posterior vitrectomy because of advanced diabetic retinopathy (tractional retinal detachment), without vitreous hemorrhage. In the case of the DME group, the indication for vitrectomy was to perform a hyaloidectomy and internal limiting membrane peeling to improve the macular thickness. None of these patients had proliferative diabetic retinopathy; therefore, there was no vitreous hemorrhage. All procedures were performed in accordance with the principles stated in the Declaration of Helsinki. All protocol procedures were approved by the Institutional Ethics Committee (approval number 19/16) and an informed consent form was signed by all participants during an interview with A.S.V. or D.R.J.
Prior to starting the vitrectomy and before opening the infusion of physiological solution, 1 mL of vitreous was cut and aspirated with the vitrector. The vitreous sample was aspirated with a sterile technique using a 1 mL syringe to extract it from the vitrector equipment hoses. The vitreous sample was placed in a sterile microtube for centrifugation and storage at −80 °C until use.
Three mL of whole blood were obtained from each participant using BD Vacutainer blood collection tubes (Becton, Dickinson, and Company, Franklin Lakes, NJ, USA), while serum was separated by centrifugation at 5000 rpm for 10 min. Total RNA was extracted from 200 μL of serum and 200 μL of vitreous samples using the miRNeasy Serum/Plasma Kit (QIAGEN, Germantown, MD, USA, Cat. No. 217184) according to the manufacturer’s instructions. RNA concentration and purity were analyzed using a BioDrop spectrophotometer (Biodrop, Cambridge, United Kingdom). RNA pools were prepared for each study group containing equal amounts of total RNA. Each pool contained total RNA from eight individuals; therefore, we generated 3 pools of total RNA derived from the serum and 3 pools of total RNA derived from the vitreous samples. The same individuals integrated the pools of serum-derived and vitreous-derived total RNA.
In each pool of RNA, quantitative global profiling of serum and vitreous miRNAs was performed using the low-density TaqMan arrays (TLDAs) Megaplex (Applied Biosystems, Foster City, CA, USA, Cat. No. 4444913) which includes panels A and B. Panel A contains 384 TaqMan microRNA assays, enabling the simultaneous quantification of 377 human mature miRNAs, in addition to 4 endogenous controls. Panel B contains 290 human mature miRNA assays, in addition to 7 endogenous controls. Differentially expressed miRNAs were identified through the Expression Suite Software (Applied Biosystems). Both panels were prepared following the manufacturer’s recommendations; briefly, 3 μL of total RNA from each pool were reverse-transcribed with the Megaplex RT primer (Applied Biosystems, California, USA, Cat. No. 4444913). The RT products were then pre-amplified using Megaplex PreAmp Primers and the TaqMan PreAmp Master Mix. The cDNA was diluted to 1:8 with distilled water and subsequently distributed into the 384 wells via centrifugation. The real-time PCR cycling parameters were set according to the manufacturer’s recommendations using a QuantStudio 7 instrument (Applied Biosystems, Foster City, CA, USA). Expression data from the cards were analyzed using the Expression Suite software (Applied Biosystems, Foster City, CA, USA), considering for expression analysis only those miRNAs with a raw Cq value below 35. The fold change was estimated using the 2−ΔΔCt method. The normalization factor was represented by the global mean expression value of all miRNAs.
The miRNAs with the most significant differential expression and published evidence regarding their potential involvement in diabetic microangiopathies were selected for further validation. Validation was carried out in the three groups of the entire population study through qPCR. For validation, 3 µL of total RNA at a concentration of 10 ng/μL was used as the template for reverse transcription, using the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA, Cat. No. 4366596). We scaled the contents of the retro-transcription reactions to obtain enough cDNA for analyzing all the genes of interest. The final volume of the RT reactions was 30 µL. Each reaction was diluted to a final volume of 100 µL with RNase-free water. Afterward, 5 µL were used as the template for qPCR reactions to analyze the expression of the selected miRNAs, using predesigned TaqMan microRNA assays (ThermoFisher, Waltham, MA, USA): hsa-miR-145-5p (Cat. No. 002278), hsa-miR-92a-3p (Cat. No. 000431), hsa-miR-375-3p (Cat. No. 000564), hsa-miR-486-5p (Cat. No. 001278), hsa-miR-212-3p (Cat. No. 000515) and hsa-miR-223-5p (Cat. No. 002098). qPCRs were performed using a QuantStudio 7 instrument (Applied Biosystems, Foster City, CA, USA). The expression levels were normalized using RNU6 as the reference gene. All the qPCR reactions were performed in triplicate. The expression fold change of all miRNAs was determined using the 2−ΔΔCt method.
Demographic features were analyzed using the IBM SPSS 15 statistics software. Student’s t-test was used for comparing two quantitative variables, a one-way ANOVA test when comparing the three groups, and a chi-square test when comparing the proportions of categorical variables. Expression Suite Software Version 1.3 (Life Technologies, Carlsbad, CA, USA) was used to assess the differential expression of miRNAs between the study groups for the TLDAs experiments. miRNA with a fold change of ≥1.5 or ≤0.7 and a p-value of <0.05 were considered for further validation. The expression levels of validated miRNAs were compared through a Kruskal–Wallis test with paired comparisons. A correlation analysis using Spearman’s test was conducted to investigate if the levels of the validated miRNAs correlated with the values of glycemia, best corrected visual acuity, and the number of years since the diagnosis of T2DM.
Target genes for those miRNAs for which differential expression was validated through qPCR were obtained using the miRNet database [14]. These target genes were used as input for pathway enrichment analysis, using WebGestalt (a web-based gene set analysis toolkit) [15]. Parameters for the analysis were over-representation analysis (ORA) being the method used, with the Kyoto Encyclopedia of Genes and Genomes (KEGG) as the functional database selected. Interaction networks between miRNAs and target genes were built on miRNet. The interaction of pathways was depicted using the enrichment map app for Cytoscape software [16,17].
Samples were successfully collected from 17 patients with DME, 16 patients with PDR, and 23 patients with IEM. The mean ages were significantly different between groups (p < 0.001). The predominant sex in the three groups was female, although the proportion difference is not significant (p = 0.614). The mean of the years since DM diagnosis was longer in the DME group compared to the PDR group (p = 0.003). The mean best corrected visual acuity (BCVA) was lower in the DME group (log-transformed in Table 1) without significance vs. controls but, when comparing both diabetic groups, patients with DME have significantly less vision than PDR (p < 0.001). About 27% of diabetic patients had been treated with an anti-VEGF more than 12 months prior to their participation in the study. The preoperative mean of glycemia is higher in the DME group (p = 0.010). Only one patient in the PDR group and two in the DME group had diabetic nephropathy. No other DM-related complications were reported in the population study. Regarding the control group comprising individuals with IEM, DM was ruled out in all of them. The mean age of the IEM group was significantly higher than the DM patients; this is related to the usual age of presentation of IEM. The average visual acuity in this group was 0.7 in logMAR, which is roughly 20/100 on Snellen’s chart (Table 1).
The expression analysis, through TLDAs, of each study group was performed by duplicate. When we compared expression profiles of patients with PDR against IEM, three miRNAs showed differential expression in the serum, hsa-miR-320a-3p, hsa-miR-92a-3p, and hsa-miR-375-3p, and two did so in the vitreous samples, hsa-miR-541-5p and hsa-miR-223-5p. When the comparison was between patients with DME against IEM, four miRNAs had significant differential expression; in the serum, these were hsa-miR-486-5p, hsa-miR-197-3p, and hsa-miR-125b-5p, while one in the vitreous samples, it was hsa-miR-212-3p. In the comparison between both groups of diabetic patients (PDR vs. DME, considering DME as the reference), five miRNAs were differentially expressed in the serum, hsa-miR-486-3p, hsa-miR-100-5p, hsa-miR-328-3p, hsa-660-5p, and hsa-145-5p. In the vitreous samples, the comparison between these two groups did not show any miRNA being differentially expressed. Table 2 summarizes these findings.
A total of 14 miRNAs showed differential expression in the TLDAs assay. To further identify the enriched functions under the regulation of the differentially expressed miRNAs, we used the target genes as the input for enrichment analysis. The analysis using miRNet revealed 1911 annotated target genes, with miRNA response elements in their 3’UTR region, for the five miRNAs that were differentially expressed in the comparison of PDR vs. IEM. For the four miRNAs with differential expression when comparing DME vs. IEM, 1417 annotated target genes were identified. These target genes were used as input for enrichment analysis. Table 3 depicts the enriched pathways from each comparison. For both comparisons, gene ontology analysis showed that the majority of the target genes participate in biological regulation, are expressed in the nucleus, and have a function related to protein binding (Figure 1).
We conducted a literature review focused on those miRNAs showing differential expression in the TLDAs analysis. Based on the previous evidence reported, four miRNAs in the serum, hsa-miR-145, hsa-miR-92a-3p, hsa-miR-486-5p, and hsa-miR-375-3p, and two in the vitreous samples, hsa-miR-223-5p and hsa-miR-212-3p, were selected for further validation through qPCR. Validation assays were carried out in all the individuals from each study group. The analysis of the four miRNAs in the serum validated the differential expression in three of them, hsa-miR-375-3p, hsa-miR-92a-3p, and hsa-miR-145, between the study groups. Further analysis revealed significant differences between all the possible pairs of groups (Figure 2). Interestingly, there were no significant differences between the IEM and DME groups. On the other hand, the expression of hsa-miR-145 and hsa-miR-92a-3p was significantly higher in the PDR group compared to the IEM group. The only miRNA with differential expression between the two groups of diabetic individuals was hsa-miR-375-3p, which showed higher expression in the DME group compared to the PDR group. We also performed qPCR validation in two miRNAs from the vitreous samples, hsa-miR-212-3p and hsa-miR-223-5p. These two miRNAs could not be detected in most of the samples of the three groups, even when the reference gene was amplified correctly. Therefore, we could not validate these two miRNAs. The correlation analysis between the three validated miRNAs and the levels of glycemia, the number of years since the diagnosis of T2DM, and the BCVA did not reveal statistically significant correlations (data not shown).
The target genes of the three validated miRNAs were obtained using the miRNet database; a total of 1326 targets were recognized. These genes were used as input for functional enrichment analysis in gProfiler for searching in the KEGG, Reactome, and Wikipathways databases. Afterward, the pathways’ interaction networks were built using the Enrichment Map app for Cytoscape. Reactome and Wikipathways revealed enrichment of the VEGF-related pathways and the closest interacting pathways were also identified (Figure 3). The focal adhesion pathway also showed enrichment in two databases, Wikipathways and KEGG.
Genetic studies are powerful tools for dissecting the molecular mechanisms underlying diabetic microangiopathies. A variety of genes have been reported to be involved in the development and progression of diabetic retinopathy [3]. Heritability has been estimated to be between 27% to 52% for these disorders [18]. The evidence demonstrating the dysregulation of miRNAs involved in DM-related metabolic pathways is constantly growing [19], and the observation that some miRNAs share mechanisms of action with multi-system diabetic microangiopathy is interesting [4]. The importance of miRNAs in regulating angiogenesis has been revealed by in vitro and in vivo studies [20]. Among the miRNAs involved in the modulation of crucial pathways for glucose metabolism, miR-375 is highly expressed in pancreatic β-cells, and is able to directly reduce insulin secretion [21,22]. We observed the upregulation of miR-375; levels of this miRNA have been found in relation to the damage at the zonula adherens and zonula occludens. The zonula occludens is closely related to the breakdown of the blood–retinal barrier. In addition, miR-375 has shown differential expression in retinal tissue derived from diabetic rats [23]. It also can regulate rat pulmonary microvascular endothelial cell activity during hypoxia by targeting Notch1 [24]. Notch1 represents a crucial pathway regulating angiogenesis in diabetic retinopathy [25]. The TLDA assay found the downregulation of miR-320 in the serum of patients with PDR, although it was not successfully validated by qPCR. In diabetic rats, miR-320 suppresses the glucose-induced increase in VEGF [26]. miR-320 has been proposed as a potential predictor of retinopathy progression in patients with type 1 DM [27]. In addition, miR-320 has been related to endothelial damage [28], targeting the inflammatory metabolic pathways. Another relevant miRNA that is downregulated in patients with PDR was hsa-miR-92a-3p; it has been described as an inhibitor of oxidative stress [20] and as a predictor of acute coronary syndrome in diabetic patients [29]. miR-92a has also been recognized as part of the endogenous miRNAs regulating the genes involved in hypertension in endothelial cells [30]. Silencing of miR-92a reduces oxidative stress and injury in diabetic nephropathy [31]. Bonauer et al. showed that hsa-miR-92a controls the growth of new blood vessels and its overexpression in endothelial cells blocks angiogenesis in vitro and in vivo [32]. miR-92a inhibits angiogenesis by targeting VEGFA and integrin subunit alpha5 [33]. miR-92a could also be part of the molecular mechanisms promoting retina neovascularization in patients with PDR. Although miR-92a seems a potential therapeutic resource, it is important to consider that even when restoring its expression for inhibiting vascularization in the retina there could be off-target effects on other tissues. miR-145 has also been associated with a senescent phenotype of smooth muscle cells (SMC) derived from patients with T2DM. This senescent phenotype can lead to DNA damage, with further vascular dysfunction [34]. Therefore, miR-145 has the potential to become a clinically useful biomarker of vascular damage. For the initial enrichment analysis, we considered the target genes of 14 miRNAs with differential expression through TLDAs. The metabolic pathways that were enriched were related to apoptosis, focal adhesion, adherens, and tight junction pathways. All these pathways are involved in pericyte loss, which is a fundamental event for the development of DR [27,34,35]. Afterward, we used the target genes of the three validated miRNAs as input for enrichment analysis. The enriched pathways were hepatocyte growth factor (HGF) and its receptor (HGFR), epidermal growth factor and its receptor (EGF and EGFR), P13K-AKt, and the tyrosine kinases receptor (RTKs) and vascular endothelial growth factor (VEGFA and VEGFR). HGF is a multifunction cytokine that plays an important role in pancreatic physiology. In terms of retinal diabetic microangiopathies, HGF seems to be involved in pericyte survival by increasing the AKT signaling pathway, leading to the strengthening of the endothelial tight junction [36]. On the other hand, the aqueous levels of HGF correlate with macular edema severity; it was increased in the cells and macrophages associated with retinal neovascularization in the murine model of an ischemic retina [37]. Several RTKs have been involved in angiogenesis, such as the EGF and its receptor, EGFR. The EGFR family of RTKs have been found to be involved in multisystemic diabetic angiopathies [38,39,40], even though EGF has been proposed as a biomarker for DR [41]. Regarding the phosphoinositide 3-kinase (PI3K)/Akt pathway, Han et al. have shown that normal vitreous has the ability to promote the growth of human pigment retinal epithelial and endothelial cells through this pathway [42]. The phosphoinositide 3-kinase (PI3K)/Akt pathway has also been linked to hyperglycemia-induced migration, proliferation, and the angiogenesis dysfunction of endothelial cells in diabetic patients and represents a growth-regulating cellular signaling pathway. They also showed that vitreous increases its own proliferation, migration, and tube formation via EGFR in human umbilical vein endothelial cells (HUVECs). On the other hand, the finding that the suppression of vitreous-induced Akt activation, cell proliferation, and migration by blocking the EGFR in cell cultures exalts the importance of the vitreous in the physiopathology of angiogenesis-related ophthalmic diseases [43]. Crucial elements for PDR development are the permeability of the blood-retinal barrier (BRB) and the importance of pericytes and endothelial cells focal adhesions (EC-FAs) as principal structural components. Focal adhesions (FAs) are composed of a high density of proteins and provide dynamic links between the extracellular matrix and intracellular cytoskeleton. Focal adhesion kinase (FAK) or vinculin is a non-receptor protein tyrosine kinase. It is principally located in the FAs and is a key molecule involved in the control of pericyte migration; this control is loss in the diabetic retina. Pericyte migration promotes increased vascular permeability and BRB leakage, which represent an early feature of PDR pathology [44,45]. Interestingly, some of the miRNAs that are most reported in the literature associated with PDR or DME did not show differential expression in our results. However, others fully coincided, and we also found miRNAs with significant expression, described for the first time in association with retinal microangiopathy. Our results add to the evidence about the role of miRNAs in the pathogenesis of DM complications affecting the retina. Nevertheless, more research is necessary to achieve a better understanding of these conditions and define the most suitable miRNAs for clinical use. We were unable to validate by qPCR those miRNAs with differential expression in vitreous samples observed through the TLDAs assays. One possible explanation is the relatively small number of individuals in each group and the low abundance of miRNA present in the vitreous samples. Further validation may require a larger sample. This study has some limitations; first, the sample size could seem relatively small. Nevertheless, the validated miRNAs regulate the target genes participating in pathways related to the physiopathology of diabetic microangiopathies. Second, we were not able to validate the miRNAs present in the vitreous samples; one possible reason could be related to the sample size, where a larger sample size could overcome the lack of detection of these miRNAs in some samples. Nevertheless, our study adds to the evidence regarding the potential role of miRNAs in ophthalmic complications in T2DM patients. Based on these and the evidence discussed above, it is tempting to assign a clinical use to the validated miRNAs. miR-375 has the potential to be a useful biomarker of disease initiation, while miR-92a has the potential to become a promising therapeutic resource, considering its effects on angiogenesis. Finally, miR-145 has the potential to be a biomarker of vascular dysfunction. All these hypotheses need to be evaluated, with a proper study design. | true | true | true |
|
PMC9601235 | Xiaoyu Cai,Meng Zhang,Fujia Ren,Weidong Fei,Xiao Zhang,Yunchun Zhao,Yao Yao,Nengming Lin | Synovial Macrophages Expression of OX40L Is Required for Follicular Helper T Cells Differentiation in the Joint Microenvironment | 21-10-2022 | follicular helper T cells,OX40L,rheumatoid arthritis,differentiation,joint microenvironment | Signaling via the OX40/OX40L axis plays a key role in CD4+ T cell development, and OX40L expression is primarily restricted to antigen-presenting cells (APCs). This study was designed to assess the role of APC-mediated OX40L expression in the context of the development of rheumatoid arthritis (RA)-associated CD4+ T cell subsets. For these analyses, clinical samples were harvested from patients with osteoarthritis and RA, with additional analyses performed using OX40−/− mice and mice harboring monocyte/macrophage-specific deletions of OX40L. Together, these analyses revealed tissue-specific roles for OX40/OX40L signaling in RA. Specifically, higher levels of synovial macrophage OX40L expression were associated with the enhanced development of T follicular helper cells in the joint microenvironment, thereby contributing to the pathogenesis of RA. This Tfh differentiation was found to be OX40/OX40L-dependent in this synovial setting. Overall, these results indicate that the expression of OX40L by synovia macrophages is necessary to support Tfh differentiation in the joint tissues, thus offering new insight regarding the etiological basis for RA progression. | Synovial Macrophages Expression of OX40L Is Required for Follicular Helper T Cells Differentiation in the Joint Microenvironment
Signaling via the OX40/OX40L axis plays a key role in CD4+ T cell development, and OX40L expression is primarily restricted to antigen-presenting cells (APCs). This study was designed to assess the role of APC-mediated OX40L expression in the context of the development of rheumatoid arthritis (RA)-associated CD4+ T cell subsets. For these analyses, clinical samples were harvested from patients with osteoarthritis and RA, with additional analyses performed using OX40−/− mice and mice harboring monocyte/macrophage-specific deletions of OX40L. Together, these analyses revealed tissue-specific roles for OX40/OX40L signaling in RA. Specifically, higher levels of synovial macrophage OX40L expression were associated with the enhanced development of T follicular helper cells in the joint microenvironment, thereby contributing to the pathogenesis of RA. This Tfh differentiation was found to be OX40/OX40L-dependent in this synovial setting. Overall, these results indicate that the expression of OX40L by synovia macrophages is necessary to support Tfh differentiation in the joint tissues, thus offering new insight regarding the etiological basis for RA progression.
Rheumatoid arthritis (RA) is a chronic form of autoimmune inflammatory disease [1]. RA affects an estimated 0.2–1% of the global population, including 0.28–0.41% of individuals in China, and ~80% of affected patients are female [2,3]. RA patients experience bone erosion and persistent inflammation of the joints characterized caused by immune cells and fibroblasts present within the synovial tissues [4]. The progressive erosion of joint tissue and consequent joint bone destruction ultimately cause the incapacitation of individuals with RA [5]. As such, further studies of the inflammatory and immune responses that occur within the synovial microenvironment in RA patients are vital in order to better guide the treatment of this debilitating disease. The T cell co-stimulatory molecule OX40 and its cognate ligand OX40L have attracted broad research interest as therapeutic targets in T cell-mediated diseases [6]. OX40/OX40L is a key regulator of both innate and adaptive immunity and is capable of regulating both macrophage and T cell function [7]. In RA, the OX40/OX40L pathway plays an important role. In RA models and RA patients, OX40 is involved in the development of RA mediated by T lymphocytes [8]. OX40L mAb administration to type II collagen (CII) immunized DBA/1 mice significantly improved disease severity [9]. T lymphocytes in synovial fluid and synovial tissue from RA patients express OX40, and OX40L is expressed on sub-lining cells in synovial tissue [9]. OX40-Fc fusion protein alleviates PD-1-Fc-exacerbated RA by suppressing the inflammatory response [10]. In addition, OX40 plays a pathogenic role in the development of autoimmune arthritis as an alternative co-stimulator of CD4+CD28− T cells, suggesting it as a potential target for immunomodulatory therapy in RA [11]. These data suggest that OX40/OX40L plays a key role in the development of RA and that the OX40/OX40L pathway in T cells may be a potential target for the treatment of RA. CD4+ T cells are key mediators of RA pathogenesis and important components of the joint microenvironment in affected patients [12]. While many researchers have focused on the specific roles that these CD4+ T cells play in RA, how interactions between these cells and synovial macrophages (SMs) contribute to RA progression remains poorly understood. Many different factors control the differentiation of CD4+ T cells, including T cell receptor signaling and major histocompatibility complex (MHC)-mediated antigen presentation [13,14]. Tumor necrosis factor (TNF) superfamily proteins are also key regulators of CD4+ T cell differentiation [6], with the OX40/OX40L signaling pathway playing a central role in the activation of particular CD4+ T cell subsets [15], as confirmed through studies conducted using OX40L−/− or OX40−/− mice [16]. OX40L expression is evident on T cells, innate lymphoid cells, and a range of antigen-presenting cell (APC) types [17]. Efforts to clarify the importance of OX40/OX40L signaling interactions between CD4+ T cells and SMs within the joint microenvironment thus have the potential to better clarify the molecular pathogenesis of RA. In the context of autoimmune disease, OX40/OX40L signaling can regulate the induction of CD4+ T cell responses, including regulatory T cells (Tregs), T follicular helper (Tfh) cells, and type 1 helper T (Th1) cells [18,19,20]. This study was designed to explore the OX40/OX40L signaling that occurs between macrophages and CD4+ T cell subsets within the joint microenvironment. Together, these analyses have the potential to offer new insight regarding the molecular etiology of RA-related joint damage.
All animal studies were performed in accordance with appropriate ethical guidelines and were approved by the Institutional Animal Care and Use Committee of Zhejiang Laboratory Animal Center (approval no. ZJCLA-IACUC-20040013). Human tissue samples were collected from patients with RA or osteoarthritis (OA) who provided full informed consent to participate. All studies involving patient samples were approved by the Ethics Committee of Women’s Hospital Zhejiang University School of Medicine (approval no. IRB-20200355-R).
Female DBA/1 mice 5–8 weeks old were purchased from GemPharmatech Co., Ltd. (Nanjing, China). OX40−/− mice were purchased from GemPharmatech Co., Ltd. (Nanjing, China). Tnfsf4fl/fl mice and B6/JGpt-Lyz2em1Cin(Cre)/Gpt mice were purchased from GemPharmatech Co., Ltd. (Nanjing, China). The Lyz2-Cre strain uses a promoter specific to myeloid cells (monocytes, mature macrophages, granulocytes) to drive codon-optimized Cre (iCre). This strain specifically expresses Cre protein in myeloid cells and the mice targeted can be used as Cre tool mice for the induction of LoxP recombination in myeloid cells. Lyz2-cre mice were bred with conditional knockout model mice to delete the gene fragment between two LoxP. Therefore, mating Tnfsf4fl/fl mice and B6/JGpt-Lyz2em1Cin(Cre)/Gpt mice results in mice with conditional deficiency of OX40L in monocytes/macrophages.
From September 2019 to August 2021, Clinical samples were obtained from 19 OA patients and 16 RA patients (patients undergoing joint replacement), including 7 and 9 with recurrent- and progressive-type disease, respectively, defined according to the clinical classification criteria of the American Rheumatism Association for knee OA [21,22]. Inclusion criteria for OA: (a) Patients diagnosed with TMJ-OA according to DC/TMD diagnostic criteria; (b) patients should be ≥18 years old at the time of signing the informed consent form, regardless of gender; (c) 18.5 kg/m2 ≤ body mass index (BMI) ≤ 35 kg/m2, and weight ≥ 50 kg for men and ≥45 kg for women; (d) no TMD-related treatment; (e) patients fully understand the purpose and requirements of the trial, voluntarily participate in the clinical trial, and sign a written informed consent. Exclusion criteria for OA: (a) Unable to walk independently, unable to participate in the study due to dysfunction; (b) knee pain caused by trauma; (c) history of knee surgery; (d) history of rheumatoid arthritis. Inclusion criteria for RA: (a) Age 18 to 65 years, regardless of gender; (b) meet the 2010 American College of Rheumatology (ACR)/European League for Rheumatology (EULAR) classification criteria with an ACR functional classification of I-III; (c) at screening, active RA is defined as at least 6/68 joints with pressure or pain on movement and at least 4/66 joints with swelling; (d) at screening, erythrocyte sedimentation rate (ESR) ≥ upper limit of normal (ULN), or C-reactive protein (CRP) > upper limit of normal (ULN); (e) body mass index [BMI = weight/height squared (kg/m2)] within the range of 18–30; (f) fully informed about the study, participate voluntarily, and have signed a written informed consent form. Exclusion criteria for RA: (a) Other types of arthritis (such as primary arthritis, post-traumatic osteoarthritis, gouty osteoarthritis, hemophilic osteoarthritis, and tuberculous arthritis); (b) bilateral knee arthroplasty (RA patients); (c) severe cardiovascular disease (such as myocardial infarction, atrial fibrillation, angina pectoris, and cardiac failure) or cerebrovascular disease (such as cerebral infarction and cerebral hemorrhage); (d) treated with bDMARD within 6 months, prolonged use of oral anticoagulant drugs (such as aspirin, warfarin, and clopidogrel). Participants’ age, smoking, BMI (body mass index), alcohol consumption, common chronic conditions (e.g., diabetes and hypertension), and drug use did not differ significantly between the RA group and osteoarthritis patients group (OA group) (Supplementary Table S1).
A murine collagen-induced arthritis (CIA) model was established as in prior reports [1]. Briefly, mice received primary and secondary immunizations on days 0 and 21, respectively. For mice in the CIA model group, these immunizations consisted of 0.1 mL of complete Freund’s adjuvant (containing 2 mg/mL chicken type II collagen (Chondrex, America, Catalog # 20011) and 4 mg/mL BCG (Hangzhou Prevention Centre, Hangzhou, China)). Beginning on day 28, two researchers independently assessed the arthritic symptoms and body weight of each mouse in this study. Arthritis severity was scored as follows: 0—normal, no joint swelling; 1—mild ankle joint or wrist swelling, or obvious swelling of the fingers; 2—moderate ankle joint or wrist swelling; 3—severe redness and swelling of the whole paw; 4—severe redness and swelling affecting more than 1 joint. Scores for each paw were summed to produce an overall score for each mouse, with a maximum possible score of 16. Mice were euthanized on day 49, at which time the spleens, hind legs, and synovial tissues of these animals were harvested for analysis.
Peripheral blood and single-cell suspensions of synovial tissues from RA patients and OA patients were used for cell sorting. The kit used for the isolation of peripheral blood mononuclear cells (PBMCs) was purchased from Miltenyi (cat. # 130-115-169). Tissue homogenizer (Servicebio, cat. # KZ-II) was used to make single-cell suspensions of synovial tissue for subsequent sorting. To ensure cell activity, we took the following measures: (1) When preparing single cells, we always ensured that the cells were in a 4 °C environment; (2) when resuspending cells, 1–2% fetal bovine serum was added to PBS; (3) when loading samples, the cells were exposed to room temperature for as little time as possible; (4) before sorting, cells were washed with EDTA-containing PBS to remove calcium ions, and DNase I (1 mg/mL) was added to remove adhesions from dead cell DNA; (5) PI staining was used to remove dead cells. Positive sorting was performed by labeling cells with appropriate surface antibodies, followed by flow cytometry-based analyses. The specimen should be fresh and should not contain more than 10% dead cells and debris. Positive sorting steps were as follows. (1) Centrifuge at 300× g for 10 min and carefully remove the supernatant as before. (2) Add antibody (10 μL/107 cells) according to instructions and incubate at 4 °C for 15 min in the dark. (3) Add 10–20 times the labeling volume of buffer, wash the cells, centrifuge at 300× g for 10 min, remove the supernatant, and repeat the wash 1 time. Add 40 μL of beads per 108 cells, add 960 μL of buffer, and incubate for 15 min at 4 °C, protected from light. Wash by centrifugation at 300× g. Remove supernatant and resuspend with buffer (500 μL/108 cells). (4) Perform magnetic sorting with a positive sorting column, using a Midi head and an LS sorting column. The magnetic beads used for this study were from Miltenyi, and included the following: Human CD19 MicroBeads (cat. # 130-050-301) (B cells), Human CD141 MicroBeads (130-090-512) (DCs), Human CD14 MicroBeads (cat. # 130-050-201) (monocytes), Human F4/80 MicroBeads (cat. # 130-110-443) (monocytes).
Flow cytometry was used to analyze sorted cells (Human CD19+ B cells, Human CD141+ DCs, Human CD14+ monocytes, Human F4/80+ monocytes) or single-cell suspension of peripheral blood mononuclear cells or single-cell suspension of synovial tissues or single-cell suspension of spleen. Tissue homogenizer (Servicebio, cat. # KZ-II) was used to make single-cell suspensions of spleen. Cells were suspended in 100 μL PBS and stained with appropriate cell surface antibodies (0.5–1.5 μL) for 30 min at 4 °C in the dark. Cells were then rinsed two times using PBS and resuspended in 100 μL of PBS for analysis with a flow cytometer. Antibodies used for this analysis included the following: Hu CD11b APC M1/70 (BD Pharmingen, Cat. # 553312), Hu CD192 BV480 LS132.1D9 (BD Pharmingen, Cat. # 747852), Hu CD11c PE B-ly6 (BD Pharmingen, Cat. # 555392), Hu CD19 FITC HIB19 (BD Pharmingen, Cat. # 555412), Hu CD14 IHC Pure M5E2 (BD Pharmingen, Cat. # 550376), Hu CD183 BV480 (BD Pharmingen, Cat. # 746283), T-BET PE 4B10 (BD Pharmingen, Cat. # 561265) (Suitable for humans and mouse), Hu CD185 Alexa (BD Pharmingen, Cat. # 558113), Hu BATF PE (BD Pharmingen, Cat. # 27120S), Hu CD4 FITC (BD Pharmingen, Cat. # 550628), Hu CD25 APC (BD Pharmingen, Cat. # 560987), Hu Foxp3 PE (BD Pharmingen, Cat. # 560046), Ms CD183 BV750 (BD Pharmingen, Cat. # 747298), Ms CD185 APC (BD Pharmingen, Cat. # 560615), BATF PE (BD Pharmingen, Cat. # 564503), Ms CD25 APC (BD Pharmingen, Cat. # 557192), Ms Foxp3 PE (BD Pharmingen, Cat. # 560408), Ms CD4 FITC (BD Pharmingen, Cat. # 553046).
The sorted cells were used for the RT-qPCR analysis. Then, total RNA was extracted from the cells for RT-qPCR experiments. In this study, Trizol was used to extract total RNA. In single-cell suspensions, Trizol preserves RNA integrity while destroying cells and lysing cellular components. After chloroform was added and centrifuged, the lysate was stratified into aqueous and organic phases, with RNA present in the aqueous phase. Total RNA was quantified by a Thermo Scientific NanodROP 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). The extracted RNA was then reverse-transcribed using the TAKARA (Japan) reverse transcription kit to synthesize cDNA. The reactions were prepared using SYBRGreen qPCR Master Mix® (TAKARA, Shiga, Japan) in a Pikoreal 96 Realtime PCR System (Thermo Scientific, USA) according to the instructions of the kit to detect the level of mRNA. The first-strand cDNA was synthesized using the Prime Script RT® kit (Takara, Shiga, Japan) according to the instructions. Reverse transcription conditions: 15 min at 37 °C, 5 s at 85 °C, 10 min at 37 °C. RT-qPCR was performed using the one-step method Thermo Step One® at 95 °C for 10 min, followed by 40 cycles at 95 °C for 15 s and 40 cycles at 60 °C, and approximately 1 min at 60 °C. The Bio-Rad PCR platform (T100PCR) was used. The number of technical replicates was 3. CT values of target and internal reference genes were obtained after real-time amplification. GAPDH is used as an internal standard. Primer information of Tnfsf4 (5’→3’): forward primer GGTCAGGTCTGTCAACTCCTT, reverse primer CATCCAGGGAGGTATTGTCAGT. Relative expression was compared using the ΔΔCT method with appropriate normalization as follows: A = CT(target gene, experimental sample) − CT(internal standard gene, experimental sample), B = CT(target gene, control sample) − CT(internal standard gene, control sample), K = A − B, relative target gene expression = 2 − K.
Hind leg and spleen tissue samples harvested from mice were fixed for 24 h in 4% paraformaldehyde. Spleen samples were then immediately paraffinized. Hind limb samples were decalcified for 1 month prior to paraffinization and isolation of the knee and ankle joints. The tissue was placed in a perforated PE tube and then placed in a decalcifying bucket, poured full of EDTA decalcifying solution, sealed, and placed in a constant temperature shaker with a decalcifying solution change cycle of 2–3 days. Paraffinized knee and ankle sections were utilized for H&E staining and safranin O-fast green staining, while spleen sections were utilized for H&E staining. Two researchers independently scored the staining results for these sections. HE staining steps were as follows. (1) Dewaxing of paraffin sections to water: sequentially put the sections into xylene Ⅰ for 20 min, xylene Ⅱ for 20 min, anhydrous ethanol Ⅰ for 5 min, anhydrous ethanol Ⅱ for 5 min, 75% ethanol for 5 min, tap water washing. (2) Hematoxylin staining: stain sections with hematoxylin solution for 3–5 min, rinse with tap water. Then, treat the section with hematoxylin differentiation solution, rinse with tap water. Treat the section with hematoxylin Scott tap bluing, rinse with tap water. (3) Eosin staining: 85% ethanol for 5 min; 95% ethanol for 5 min. Finally stain sections with eosin dye for 5 min. (4) Dehydration and sealing: dehydrate as follows: 100% ethanol I for 5 min, 100% ethanol II for 5 min, 100% ethanol III for 5 min, xylene I for 5 min, xylene II for 5 min, and finally seal with neutral gum. (5) Microscopic examination, image acquisition and analysis. Safranin O-fast green staining steps were as follows. (1) Paraffin sections dewaxed to water: sequentially put the sections into environmentally friendly dewaxing transparent solution Ⅰ for 20 min, environmentally friendly dewaxing transparent solution Ⅱ for 20 min, anhydrous ethanol Ⅰ for 5 min, anhydrous ethanol Ⅱ for 5 min, 75% ethanol for 5 min, and wash with tap water. (2) Fast green staining: section into bone tissue solid green staining solution for 1–5 min, wash with water to remove excess staining solution until the cartilage is colorless, soak in 1% hydrochloric acid ethanol for 10 s, wash slightly with tap water. (3) Saffron staining: the slides were stained in saffron dye solution for 1–5 s, and then put into four cylinders of absolute ethanol, for rapid dehydration for 5 s, 2 s, and 10 s, and kept in the fourth cylinder. (4) Transparent sealing: clean xylene transparent 5 min, neutral resin sealing. (5) Microscopic examination, image acquisition and analysis. H&E staining scoring criteria for knee and ankle samples were based on the degree of synovial cell hyperplasia (0, no hyperplasia; 1, mild hyperplasia; 2, moderate hyperplasia; 3, severe hyperplasia), the degree of vascular hyperplasia (0, no vascular hyperplasia; 1, mild vascular hyperplasia; 2 moderate vascular hyperplasia; 3 severe vascular hyperplasia), the degree of fibrous tissue hyperplasia (0, no fibrous tissue hyperplasia; 1 mild fibrous tissue hyperplasia; 2, moderate fibrous tissue hyperplasia; 3, severe fibrous tissue hyperplasia), and the degree of lymphocytic infiltration (0, no infiltration; 1 mild infiltration; 2, moderate infiltration; 3 severe infiltration). Safranin O-fast green staining of ankle and knee sections was used to visualize bone tissue, articular cartilage, and subchondral bone structures, with adult bone and cartilage stained green and red, respectively. H&E staining of spleen samples was used to assess the numbers and sizes of germinal centers in the spleen. Scoring was performed by two independent researchers (skilled in scoring joint HE staining), and the results were taken as the mean value.
Data are means ± standard deviation (SD) and were compared using Student’s t test (two groups) or one-way ANOVA (three or more groups) after meeting the requisite assumptions for these statistical tests. Numbers of mice in individual experiments are indicated in the figure legends, and all mouse group assignments were random. A blinded approach was used when scoring clinical and histological samples. p < 0.05 was the significance threshold, and all analyses were performed using GraphPad Prism (v 8.01) for Windows.
For this study, we collected samples of peripheral blood and synovial tissue from 16 RA patients (seven recurrent type and nine progressive type) and 19 OA patients according to inclusion and exclusion criteria. Initially, we analyzed OX40L expression in synovial APCs from these samples by using positive bead-based sorting to isolate single-cell suspensions of F4/80+ SMs, CD19+ B cells, and CD141+ dendritic cells (DCs). Relative to corresponding cell populations from patients with OA, OX40L expression was significantly increased in synovial tissue F4/80+ SMs (Figure 1A), CD19+ B cells (Figure 1B), and CD141+ DCs (Figure 1C) from RA patients. Flow cytometry also revealed a higher frequency of CD11b+CD192+ SMs (Figure 1D), CD19+ B cells (Figure 1E), and CD11b+CD11c+ DCs (Figure 1F) in synovial tissue samples from individuals with RA relative to individuals with OA. To extend these analyses further, we sorted monocytes (CD14+), B cells (CD19+), and DCs (CD141+) from patient PBMCs. As in synovial tissues, significantly a higher level of OX40L expression was detected in the monocytes (Figure 1G), DCs (Figure 1H), and B cells (Figure 1I) of RA patients relative to OA patients, and RA patients also exhibited a higher frequency of circulating CD11c+ DCs (Figure 1J), CD19+ B cells (Figure 1K), and CD14+ monocytes (Figure 1L) as compared to OA patients. Notably, OX40L in SMs only showed an increase in synovial tissues of RA patients compared to peripheral blood (Figure 1O). Furthermore, the relationship between OX40L level and DAS28 scores (Figure 1M) was analyzed. Interestingly, the OX40L level in CD11b+CD192+ SMs in synovial tissues was positively correlated with DAS28 (Figure 1N), whereas other cells were not (Supplementary Figure S1A–E). These data may suggest that OX40L plays a tissue-specific role in the synovium in individuals with RA, shaping optimal T cell activation in this compartment. These results also support a relationship between OX40L expression and joint damage severity in RA patients.
To better understand the association between the expression of OX40L by SMs and RA-related disease progression, we next analyzed three different CD4+ T cell subsets that are closely linked to OX40/OX40L signaling activity (Th1, Tregs, and Tfh) [19,23,24], using a flow cytometry-based approach to assess the relative frequencies of these three different subsets as a percentage of the overall CD4+ T cell population in RA patient synovial tissue samples. Relative to synovial tissue samples from OA patients, those from RA patients contained increased percentages of CD183+T-bet+ Th1 (Figure 2A) and CD185+BATF+ Tfh (Figure 3B) cells, whereas the frequency of CD25+Foxp3+ Tregs (Figure 2C) was reduced. Notably, a positive correlation was observed between OX40L level in SMs and CD185+BATF+ Tfh frequency (Figure 2D), whereas no such correlation was observed for CD183+T-bet+ Th1 (Figure 2E) or CD25+Foxp3+ Tregs (Figure 2F) in RA patients. These data indicate that increased OX40L expression by SMs may play a role in promoting Tfh differentiation in a manner that ultimately promotes increased RA disease activity.
Many different factors can shape CD4+ T cell activation and differentiation, including both TCR and OX40/OX40L signaling [24,25,26]. To further explore the importance of OX40/OX40L signaling in the pathogenesis of RA, we established a collagen-induced arthritis (CIA) model by using OX40−/− mice. Strikingly, OX40−/− mice exhibited significantly decreased arthritic disease severity as compared to wild-type (WT) animals (Figure 3A). This coincided with the alleviation of knee pathology in these animals, with lower levels of bone damage, synovial hyperplasia, and inflammatory cell infiltration (Figure 3B,C). Specifically, the wild-type mice had severe cartilage damage, marked synovial hyperplasia, and multiple inflammatory cell infiltrates in the joint cavity, whereas the OX40−/− mice had less cartilage damage, less marked synovial hyperplasia, and a few inflammatory cell infiltrates in the joint cavity. Synovial tissue single-cell suspensions from OX40−/− mice revealed slight decreases in the frequency of CD183+T-bet+ Th1 as a fraction of total CD4+ T cells in the synovial tissues relative to WT mice (Supplementary Figure S2A), with a corresponding increase in CD25+Foxp3+ Treg frequency (Supplementary Figure S2B). Notably, no significant differences in CD183+T-bet+ Th1 (Supplementary Figure S2C) or CD25+Foxp3+ Treg (Supplementary Figure S2D) frequencies were observed in the peripheral blood relative to the synovial tissues (Figure 3D). Compared with wild-type mice, the frequency of CD4+Tfh in the synovium of OX40−/− mice decreased by 84.8 ± 1.5% (Figure 3E), while the frequency of CD4+ Tfh in the peripheral blood decreased by 30.8 ± 1.1% (Figure 3F). The percentage of decrease in the frequency of CD4+ Tfh in the synovium of OX40−/− mice was significantly different from that in the peripheral blood (Figure 3G). These results align well with findings from RA patients detailed above, suggesting that OX40/OX40L signaling plays a tissue-specific role in RA, supporting Tfh differentiation within the joint microenvironment in arthritic model mice.
As OX40L expression by SMs was positively correlated with DAS28 values and with the frequency of Tfh as a fraction of the overall CD4+ T cell population, the SM OX40L expression level may serve as a key driver of Tfh differentiation in the RA-associated joint microenvironment. To test this possibility, we employed mice exhibiting macrophage-specific OX40L deletion (Tnfsf4fl/fl/Lyz2-Cre mice) and control Tnfsf4fl/fl mice to establish a CIA model as above. After 49 days, these mice were euthanized for downstream analyses. Tnfsf4fl/fl/Lyz2-Cre mice exhibited significant reductions in paw swelling (Figure 4A), inflammatory cell infiltration (Figure 4B), and bone damage (Figure 4C) in the knee joint. Specifically, the Tnfsf4fl/fl mice had severe cartilage damage, marked synovial hyperplasia, and multiple inflammatory cell infiltrates in the joint cavity, whereas the Tnfsf4fl/fl/Lyz2-Cre mice had less cartilage damage, less marked synovial hyperplasia, and a few inflammatory cell infiltrates in the joint cavity. Compared to Tnfsf4fl/fl mice, the frequency of CD4+ Tfh was down-regulated in the knee cavity of Tnfsf4fl/fl/Lyz2-Cre mice (Figure 4D), whereas no corresponding differences in the frequencies of CD183+T-bet+ Th1 (Figure 4E) or CD25+Foxp3+ Tregs (Figure 4F) were observed. These findings are consistent with a model wherein the expression of OX40L by SMs within the knee microenvironment is required for Tfh differentiation in this mouse CIA model system, whereas this OX40L signaling is dispensable for Th1 or Treg differentiation.
To further explore the tissue-specific nature of the impact of OX40L expression on SMs in the context of Tfh differentiation, we analyzed PBMCs from these experimental mice. No differences in the frequencies of CD185+BATF+ Tfh were observed as a fraction of total CD4+ T cells in Tnfsf4fl/fl/Lyz2-Cre mice compared to Tnfsf4fl/fl mice (Figure 5A). Similarly, there were no significant differences in the proportions of CD183+T-bet+ Th1 (Figure 5B) or CD25+Foxp3+ Tregs (Figure 5C) as a fraction of total CD4+ T cells. These results suggest that the expression of OX40L on monocytes is dispensable for the differentiation of Tfh in the peripheral blood. While other factors can evidently support Tfh development in the periphery, these same signals fail to facilitate such differentiation in the joint microenvironment in the absence of OX40L.
We additionally analyzed CD4+ T cell and CD19+ B cell subsets in the spleens of these mice. Relative to Tnfsf4fl/fl mice, no significant differences were observed in Tnfsf4fl/fl/Lyz2-Cre mice with respect to germinal center size (Figure 6A), CD19+CD23+CD24+ transitional B cell frequency (Figure 6B), or follicular B cell (CD19+CD23+CD38+) frequency (Figure 6C). Consistently, there was no difference in the proportion of CD185+BATF+ Tfh (Figure 6D), CD183+T-bet+ Th1 (Figure 6E), or CD25+Foxp3+ Tregs (Figure 6F) among CD4+ T cells in the spleen of Tnfsf4fl/fl/Lyz2-Cre mice. These findings indicate that Tfh, Th1, and Treg differentiation does not require OX40L expression by splenic macrophages, supporting the tissue-specific importance of monocyte OX40L expression for Tfh differentiation.
Synovial macrophages are essential mediators of the inflammatory activity and bone damage observed in the joints of RA patients [27]. These SM populations are highly heterogeneous such that researchers have defined particular SM subtypes within the synovium [28]. The present results suggest that the expression of OX40L by SMs is vital to the effective development of Tfh within the joint microenvironment in the context of RA. The initial analyses in this study were centered around synovial tissue samples from individuals with RA and OA based on the rationale that all medically focused research should be of clinical origin and clinically relevant where possible. OA patients served as controls in this study owing to the challenges associated with collecting synovial tissue samples from individuals without any form of joint disease. OX40L expression on the three most abundant APC populations (DCs, B cells, and monocytes/macrophages) was then analyzed in both the synovium and peripheral blood of these patients. In line with prior work, all three of these cell populations were found to exhibit increased OX40L expression in both the peripheral blood and synovium relative to OA patients. Strikingly, only OX40L expression by SMs was positively correlated with RA patient disease activity scores. These results suggest the possibility that OX40L may play a more localized role in the pathogenesis of RA, rather than shaping this disease at the systemic level. The OX40L-mediated regulation of CD4+ T cell activation may thus be a key pathway through which OX40/OX40L signaling can control RA development. To gain further insight into the observed correlation between OX40L expression by SMs and DAS28 scores, the three CD4+ T cell subsets most closely related to OX40/OX40L signaling (Tfh, Th1, and Tregs) were examined, given that they can control RA-related inflammatory activity and immune responses [6,8,19,29]. These analyses revealed a positive correlation between the frequency of CD185+BATF+ Tfh as a fraction of total CD4+ T cells and SM OX40L expression, whereas the same correlative relationship was not detected in peripheral blood samples. This suggested the ability of OX40L expressed by SMs to control inflammation and immune response activation within the joint microenvironment through the regulation of Tfh differentiation in a tissue-specific manner. This model was further supported by the use of OX40−/− mice, which revealed that while the impact of losing OX40 expression was systemic in a CIA model system, the corresponding impact on Tfh differentiation was tissue-specific. Specifically, the differentiation of Tfh was found to be OX40/OX40L signaling-dependent. When this signaling was no longer available, compensatory mechanisms were sufficient to sustain Tfh differentiation in the peripheral blood but not within the joint microenvironment. To further explore the importance of OX40L expression by SMs in the context of Tfh differentiation, macrophage-specific deletion of OX40L in mice (Tnfsf4fl/fl/Lyz2-Cre mice) and control mice (Tnfsf4fl/fl mice) was used in the study. Tnfsf4 is the gene encoding the OX40L protein, and Lyz2 is predominantly expressed in macrophages. Lyz2-Cre mice were mated with Tnfsf4fl/fl mice to obtain Tnfsf4fl/fl/Lyz2-Cre mice. Our data revealed that differentiation of Tfh requires SMs to express OX40L in the articular microenvironment. Through gene editing studies in mice, our data reveal that the differentiation of Tfh requires SMs to express OX40L in the joint microenvironment. However, this phenomenon was not observed in the peripheral blood and spleen of Tnfsf4fl/fl/Lyz2-Cre mice, which laterally corroborates that the differentiation of Tfh is tissue-specific with respect to the requirement for OX40L. Tfh are essential mediators of B cell activation and differentiation, promoting germinal center formation and immunoglobulin class switching such that they are closely related to the development of humoral immune responses [30,31]. Tfh differentiation occurs through pathways distinct from those employed by other related CD4+ T cell populations such as Th1, Th17, and Tregs [32,33]. The process of Tfh differentiation entails initiation, maintenance, and polarization phases that rely on the coordinated activity of many surface molecules, cytokines, and transcriptional regulators [26]. During the initiation phase, Tfh must interact with APCs. In line with these prior reports, the present data emphasize the fact that Tfh differentiation is distinct from that of Th1 and Tregs. There are certain limitations to these analyses. For one, only specific APC subsets and CD4+ T cell subpopulations were analyzed, potentially resulting in phenotypes associated with other similar cell types having been overlooked. As such, further research will be essential to fully clarify the tissue-specific roles of OX40/OX40L signaling.
In summary, these results indicate that OX40L is upregulated in APCs isolated from RA patient synovial tissues, and that the expression of OX40L by CD11b+CD192+ SMs is positively correlated with RA disease severity. The OX40L level in SMs is also positively associated with Tfh frequency in RA patient synovial tissues, while in CIA model mice, the differentiation of Tfh within the joint microenvironment is dependent on OX40/OX40L signaling. In contrast, Tfh, Th1, and Treg differentiation in the peripheral blood of CIA model mice is not dependent on OX40L expression by SMs or splenic OX40L expression. Overall, these data highlight the importance of OX40L expression by SMs as an essential mediator of Tfh development within the RA-associated joint microenvironment. | true | true | true |
|
PMC9601708 | Qiang Ding,Xiuhu Ding,Shuwen Xia,Fang Zhao,Kunlin Chen,Yong Qian,Shaoxian Cao,Zhiping Lin,Yundong Gao,Huili Wang,Jifeng Zhong | Bta-miR-6531 Regulates Calcium Influx in Bovine Leydig Cells and Is Associated with Sperm Motility | 03-10-2022 | sperm motility,bta-miR-6531,ATP2A2,calcium signaling pathway,SNP | MicroRNAs (miRNAs) play key roles in sperm as the regulatory factors involved in fertility and subsequent early embryonic development. Bta-miR-6531 is specifically a highly enriched miRNA in low-motility sperms in previous study. To investigate the mechanism of bta-miR-6531, 508 shared target genes of bta-miR-6531 were predicted using two miRNA target databases (TargetScan7 and miRWalk). According to the Kyoto Encyclopedia of Genes and Genomes (KEGG), the calcium and cAMP signaling pathways were the most enriched of the target genes. A dual-luciferase assay indicated that bta-miR-6531 targeted ATP2A2 mRNA by binding to the coding sequence region. In bovine Leydig cells, bta-miR-6531 overexpression affected the intracellular calcium concentration by restraining ATP2A2 expression. Moreover, we observed high calcium concentrations and high ATP2A2 protein levels in high-motility sperm compared with those in low-motility sperms. Furthermore, high-linkage single-nucleotide polymorphisms (SNPs) of the pre-bta-miR-6531 sequence were identified that related to sperm traits. Genotype TCTC of bta-miR-6531 showed high sperm motility and density and low deformity rate in Holstein bulls. However, the mutation in pre-miR-6531 did not significantly affect mature bta-miR-6531 expression in the sperm or cell models. Our results demonstrate that bta-miR-6531 might involve in sperm motility regulation by targeting ATP2A2 of the calcium signaling pathway in bovine spermatozoa. | Bta-miR-6531 Regulates Calcium Influx in Bovine Leydig Cells and Is Associated with Sperm Motility
MicroRNAs (miRNAs) play key roles in sperm as the regulatory factors involved in fertility and subsequent early embryonic development. Bta-miR-6531 is specifically a highly enriched miRNA in low-motility sperms in previous study. To investigate the mechanism of bta-miR-6531, 508 shared target genes of bta-miR-6531 were predicted using two miRNA target databases (TargetScan7 and miRWalk). According to the Kyoto Encyclopedia of Genes and Genomes (KEGG), the calcium and cAMP signaling pathways were the most enriched of the target genes. A dual-luciferase assay indicated that bta-miR-6531 targeted ATP2A2 mRNA by binding to the coding sequence region. In bovine Leydig cells, bta-miR-6531 overexpression affected the intracellular calcium concentration by restraining ATP2A2 expression. Moreover, we observed high calcium concentrations and high ATP2A2 protein levels in high-motility sperm compared with those in low-motility sperms. Furthermore, high-linkage single-nucleotide polymorphisms (SNPs) of the pre-bta-miR-6531 sequence were identified that related to sperm traits. Genotype TCTC of bta-miR-6531 showed high sperm motility and density and low deformity rate in Holstein bulls. However, the mutation in pre-miR-6531 did not significantly affect mature bta-miR-6531 expression in the sperm or cell models. Our results demonstrate that bta-miR-6531 might involve in sperm motility regulation by targeting ATP2A2 of the calcium signaling pathway in bovine spermatozoa.
MicroRNAs (miRNAs), a class of short single-stranded non-coding RNAs of with a length of approximately 18–24 nt, are known to negatively regulate mRNA expression or repress protein translation by binding to mRNA transcripts [1]. Several studies have reported that sperm miRNAs play essential roles in the fertilization of zygotes and in early embryonic development [2,3,4]. Deep sequencing has identified thousands of miRNAs that are differentially expressed between low- and high-motility sperm in cattle [2,3,5]. Canonical biogenesis pathways of miRNAs have been defined. In the nucleus, miRNAs are initially transcribed from long primary transcripts (termed pri-miRNAs) and further processed into precursor miRNAs (pre-miRNAs) by RNase Drosha [6]. Then, the pre-miRNAs are exported to the cytoplasm by Exportin 5 and further cleaved into mature miRNAs by DICER [1]. Mature miRNAs recognize their target mRNAs mainly by the base-pairing interactions between nucleotides 2–8 (seed region) from the 5′ end and complementary nucleotides in the 3′ untranslated region of the target gene mRNA. Thus, it is estimated that a miRNA may target hundreds of genes in the mammalian transcriptome and many miRNAs regulated most genes. Sperm quality is the most important index in the selection of breeding bulls. Sperm quality traits are evaluated using several parameters, including sperm motility, density, viability, and morphology. Sperm motility is a major trait with high heritability (close to 0.60) in bull breeding programs [7]. With the development of molecular marker techniques and marker-assisted selection in animal breeding, single-nucleotide polymorphisms (SNPs) in candidate genes could be used as markers to predict sperm quality traits in bulls. Over the past two decades, the number of SNPs in quantitative trait loci (QTL) has been related to sperm quality in cattle [7,8]. However, the genetic potential of the reproductive characteristic of Holstein bulls has not been fully explored. miRNA-related SNPs have been demonstrated to have complex effects by affecting the miRNA function, including miRNA maturation, functional strand selection, and target gene selection [9]. Mutations in the miRNA precursor sequences potentially cause mammalian diseases and influence biological traits [10,11]. Based on our previous findings, we used a customized miRNAQTLsnp software to identify several SNPs loci in bovine miRNA such as pre-bta-miR-6531 [12]. These SNPs were also found to be co-located at the quantitative trait loci (QTLs) that may be associated with semen quality. Therefore, in the present study, we aimed to elucidate the molecular mechanism of miR-6531 regulation of sperm motility in cattle, and further confirm whether the candidate SNPs in the bta-miR-6531 precursor sequence are related to sperm quality.
Animals were not used in this study. All Bovine semen was provided by the Shandong OX Livestock Breeding Co., Ltd. (Jinan, China).
Based on the bta-miR-6531 sequence provided by the miRbase [13], two miRNA databases containing the cow database, TargetScan (http://www.targetscan.org/, accessed on 29 September 2021) [14] and miRWalk (http://mirwalk.umm.uni-heidelberg.de/, accessed on 29 September 2021) [15], were used to predict bta-miR-6531 target genes. The shared genes of the two databases were further used to perform KEGG pathways enrichment analyses using KOBAS software (KEGG Orthology-Based Annotation System [http://kobas.cbi.pku.edu.cn/, accessed on 29 September 2021]). Significantly enriched pathways were identified using hypergeometric distribution and Fisher’s exact test at 5% level of significance (p ≤ 0.05).
The target gene ATP2A2 CDS region sequences harboring the binding seed region of bta-miR-6531 were amplified using primers containing two restriction sites, which are listed in Supplementary Table S1. PCR products were isolated and inserted into the pmirGLO vector, following the T4 ligase according to the manufacturer’s instructions. Mutations in the seed regions were constructed via site-directed mutagenesis using the Mut Express II Fast Mutagenesis Kit V2 (C214-01, Nanjing Vazyme Biotech Co., Nanjing, China). HeLa cells were seeded in a 12-well plate and cultured in DMEM/F12 (SH30023.01B, HyClone, Logan, UT, USA) medium supplemented with 100 U/mL penicillin, 100 µg/mL streptomycin (ThermoFisher), and 10% fetal bovine serum (FBS [HyClone, Marlborough, MA, USA]) at 37 °C and 5% CO2. After 24 h, when adherent cells reached 70% confluence, the pmirGLO-targets-3′UTR or mutagenesis vectors (400 ng), synchronically with miR-6531 mimics or negative control, were co-transfected into cells according to the Lipofectamine3000 (Invitrogen, Waltham, MA, USA) protocol. After 48 h, the dual-luciferase assay system (Promega, corporation, Madison, WI, USA) was used according to the manufacturer’s instructions. Cells were lysed and fluorescence intensity was detected using a multifunctional microplate reader (GLOMaX-Muti Plus [Promega Corporation, Madison, WI, USA]). Firefly luciferase activity was normalized to Renilla luciferase activity in each transfected well. All experiments were repeated three times.
Leydig cells from bovine testes were separated using density gradient centrifugation as previously reported [16]. Briefly, testis tissues were digested with collagenase IV at 37 °C for 30 min. The digestion mixture was centrifuged at 800× g for 5 min, and the supernatant was discarded to remove the cellular debris. Furthermore, the cell pellet was layered on different Percoll density gradients (70%, 55%, and 35%) and centrifuged at 2000× g for 1 h at 4 °C. The cells in 55% Percoll were collected and resuspended in Dulbelcco`s Phosphate Buffer Saline (HyClone, Logan, UT, USA) to remove the remaining Percoll. Cells were cultured in DMEM/F12 medium supplemented with L-glutamine, 10% FBS and nonessential amino acids at 34 °C and 5% CO2. The purity of the Leydig cells was determined using the immunohistochemical of the 3 β-hydroxysteroid dehydrogenase (3β-HSD).
The fluorescent calcium chelating dye Fluo-4 AM was used as an indicator of the relative levels of intracellular calcium in Leydig cells transfected with NC mimics or bta-miR-6531 mimics in Confocal Dishes (NEST). Cells were incubated with Fluo-4 AM (2 μM) diluted in Hank’s balanced salt solution (HBSS, calcium, and magnesium free, HyClone) for 30 min at room temperature. After Fluo-4 AM was loaded, the cells were washed with HBSS buffer to remove residual Fluo-4 AM. The fluorescence of calcium was captured by a fluorescence microscopy at 100 ms exposure. The intensity of calcium signaling was measured by using Image Pro-Plus 6.0 and remove the background. For flow cytometric analysis of calcium, cells were digested by 0.25% trypsin and resuspend before load Flou-4 AM. Cell calcium flux were stimulated by ionomycin (2 μM).
The transcriptomes of sperms and cells were isolated by using TRIzol reagent (15596018, Ambion Inc., Austin, TX, USA) according to the protocol of the PureLink RNA Mini Kit (12183018A, Ambion Inc.). RNA extraction from sperms were performed using guanidinium thiocyanate supplemented with Tris(2-carboxyethyl)-phosphine (646547-10 × 1 mL, Sigma-Aldrich) [17]. For mRNA expression analysis, first strand cDNA was synthesized using HiScript® III 1st Strand cDNA Synthesis Kit (R111-01, Nanjing Vazyme Biotech Co.) with random hexa-primers. qPCR reactions were performed using ChamQ SYBR Color qPCR Master Mix (Q411-02, Nanjing Vazyme Biotech Co.) on QuantStudio 5 (Thermo Fisher Scientific, Waltham, MA, USA). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a reference gene to normalize reactions in mRNA analysis. The relative expression of each sample was calculated using the 2−ΔΔCt method [18]. The primers used for qPCR are listed in Supplementary Table S1.
For protein extraction, cells were incubated in radioimmunoprecipitation assay (RIPA) lysis buffer (P0013B, Beyotime Biotechnology, Shanghai, China) supplemented with a protease and phosphatase inhibitor cocktail (P1051, 50×, Beyotime Biotechnology) and placed on ice for 30 min. For sperm protein analysis, the sperm samples were selected based on different motility (Table S2). 1% Sodium dodecyl sulfate was added to the RIPA lysis buffer. Sperm samples were placed on ice under ultrasonic conditions for 10 cycles of 5 s pulse with a 30 s interval. The protein concentrations of all samples were measured by using the BCA assay (P0010S, Beyotime Biotechnology). Protein samples were mixed with Sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer (P0015L, Beyotime Biotechnology) and boiled for 10 min. Additionally, 20 μg total proteins was loaded onto a 12% SDS-PAGE gel for separation based on the protein size. The proteins were then transferred onto a polyvinylidene difluoride (PVDF) membrane (Roche, Mannhein Germany) at 260 mA for 1 h. The membranes for immunoblotting were blocked with 5% non-fat milk powder at room temperature for 2 h and incubated with primary antibodies (ATP2A2, 1:500; ATP1A1, 1:500; β-actin, 1:2000) at 4 °C overnight. Secondary antibodies (anti-mouse 1:2000 or anti-rabbit 1:2000) were used depending on the primary antibody. β-actin was used as a reference protein to measure the relative expression of target proteins.
The Holstein bulls in the present study ranged between 2.5 and 11 years old and were obtained from the Shandong bull station (Jinan, China). Fresh semen samples were collected using a bovine artificial vagina, quickly transferred to the laboratory, and incubated at 37 °C after collection from each bull. Subsequently, the semen volume, sperm concentration and sperm motility were determined. Ejaculate volume was calculated using a collecting vial, the density of fresh sperms was determined using a hemocytometer. Semen was packaged in 0.25 mL straws after diluting with Bioxcell (IMV Biotechnology, Basse-Normandie, France) and cryopreserved. After storage in liquid nitrogen 7 days, 2 straws were randomly selected from each ejaculate, thawed at 38 °C for 20 s, and immediately evaluated for sperm motility and deformity rate. For motility of fresh or post-thawed sperm, a drop of semen was placed on a pre-warmed (37 °C) glass slide and covered with a glass slip on a thermo plate. The sperm motility was evaluated by using a phase-contrast microscope at 400× g magnification on a TV monitor with a sperm analysis system (AndroVision Minitube, Germany). For sperm motility calculate, all processive movement or quickly movement sperms were considered as motile sperms in this study. The percentage of strawed sperm deformities was evaluated using Giemsa staining solution (Beijing Solarbio Science & Technology Co., Ltd., Beijing, China) at 400× g and 1000× g magnification [19], with more than 200 stained sperms in consideration. Some morphology defects, coiled or hairpin or terminal droplet tail, detached head and no or small acrosome were considered as abnormal sperm in the present paper.
Our previous study identified the sequence of pri-miR-6531 containing SNPs located in QTLs associated with sperm motility in Holstein bulls. Therefore, we amplified the pri-miR-6531 sequence and performed Sanger sequencing. Fifty-six bulls with semen parameter records were used in the association analysis of the SNPs of pri-miR-6531 and semen quality traits. Semen quality traits were collected and provided by Shandong OX Animal Breeding Co., Ltd., Shandong, China. The genomic DNA of each sperm sample was isolated from the frozen semen according to a previous report [8]. The pair of specific primers for the pri-miR-6531 sequence is listed in Supplementary Table S1.
The structures of the miRNA precursors were determined by using mRNA structure software (RNAstructure, Version 6.4, Developed by Jessica S. Reuter & Maintained by Richard M. Watson, New York, USA). For analysis of the bta-miR-6531 expression level, cDNA and quantitative PCR (qPCR) were performed using miRNA 1st Strand cDNA Synthesis Kit (by stem-loop [MR101-01, Nanjing Vazyme Biotech Co., Nanjing, China]) and miRNA Universal SYBR qPCR Master Mix (MQ101, Nanjing Vazyme Biotech Co., Nanjing, China) according to the manufacturer’s instructions. U6 was used as the reference miRNA for normalization, and the fold change relative to the control samples was determined by the 2−ΔΔCt method. All primer pair sequences are listed in Supplementary Table S1.
All data are shown as the means ± SEM. Experiments for each group were repeated at least three times. A two-tailed Student’s t-test was used for group comparisons. Three or more groups were compared using one-way analysis of variance and Tukey’s test. All statistical graphics were drawn in GraphPad Prism 9 software. Differences of p < 0.05 and differences of p < 0.01 were considered statistically significant and highly statistically significant, respectively.
To further explore the expression of bta-miR-6531 in bovine sperm, sperm with high or low motility were separated. qPCR analysis showed that bta-miR-6531 was lost in sperm with high motility (Figure 1A). To understand the function of bta-miR-6531, we evaluated its cellular targets. Thus, 2349 and 2739 target genes were predicted using miRWalk and TargetScan 7, respectively. Among all the target genes, 508 were shared in the two miRNA databases (Figure 1C). The KEGG pathway of shared targets genes showed that the most enriched pathways were axon guidance, cAMP signaling pathway, calcium signaling pathway, and thyroid hormone signaling pathway, all of which are actively involved in sperm (Figure 1B).
Pathway enrichment analysis showed that bta-miR-6531 directly targeted the calcium signaling pathway and the cAMP signaling pathway. We further performed intracellular calcium concentration [Ca2+]i assays and mitochondrial membrane potential (MMP) in Leydig cells. The fluorescence intensity of [Ca2+]i in bta-miR-6531-overexpressing cells showed a low concentration of [Ca2+]i compared with NC groups (Figure 2A). The results of flow cytometry of the [Ca2+]i signaling showed a low level after overexpressing bta-miR-6531, but present same level after stimulated by ionomycin (Figure 2B). Additionally, overexpression of bta-miR-6531 enhanced the MMP in Leydig cells (Figure S1), indicating that bta-miR-6531 plays a biological role in regulating calcium in bovine Leydig cells.
Bioinformatic analysis showed that bta-miR-6531 targets the ATP2A2 gene and identifies the existence of a putative bta-miR-6531 binding site within the ATP2A2 CDS region. To confirm whether bta-miR-6531 directly binds to the ATP2A2 CDS region, luciferase reporter vectors expressing the bta-miR-6531 binding site in the ATP2A2 CDS region or mutated binding site were transfected into HeLa cells and the relative luciferase activity was assessed (Figure 3A). The relative luciferase activity of wild-type reporters was significantly reduced in cells co-transfected with bta-miR-6531 mimics compared with that of NC mimics, whereas the mutant reporter vector was unaffected (Figure 3B). To confirm the target genes affected by bta-miR-6531, mRNA and protein expression were detected in Leydig cells after overexpression or interference of miR-6531 (Figure 3C). In the Leydig cell model, overexpression of bta-miR-6531 did not markedly affect the mRNA expression of target genes; however, we observed that the ATP2A2 protein was significantly upregulated after the overexpression of miR-6531 compared with that of NC mimics. Contrastingly, the inhibition of miR-6531 expression resulted in a very low ATP2A2 protein expression level compared with that of the NC inhibitor (Figure 3C).
To investigate the calcium flux in cells whether affect by ATP2A2, we designed small interfere RNA (siRNA) of target CDS region of bta-miR-6531. The results showed that the Ca intensities were lower in cells after knockdown the ATP2A2 expression than negative control (Figure 4A). To further explore the relationship between miR-6531 and the expression of its target genes in cattle sperm, we extracted proteins from sperms with high or low motility. The Western blotting results showed that ATP2A2 protein was significantly highly accumulated in sperm with high-motility than in low motility (Figure 4B). The above results demonstrated that ATP2A2 may involve in regulate sperm motility by regulate calcium influx.
PCR products were sequenced using specific primers to align the sequences of the bta-miR-6531 precursor. SNP data were created from sperm DNA samples from 56 bulls by Sanger sequencing. Sequence alignment revealed four SNPs (A-T, T-C, C-T, and T-C) located on the sequence of the bta-miR-6531 precursor, specifically, downstream of the mature sequence of bta-miR-6531 (Figure 5A). The structure of pre-miR-6531 was predicted by RNA fold [20]. The results showed that mutations in the precursor did not considerably affect the pre-miR-6531 structure (Figure 5B). Furthermore, the locations of the four SNPs at the genomic level were close, and linkage disequilibrium analysis showed that they were strongly linked (Figure 5C). To identify whether mutations within the precursor of bta-miR-6531 affect mature miRNA expression, we constructed two genotypes of miRNA precursor overexpression vectors and transfected them into bovine Leydig cells. qPCR results showed that mutation of the miR-6531 precursor did not markedly influence mature bta-miR-6531 expression (Figure 5D), indicating that the mutation of pre-bta-miR-6531 is not a functional change but could be used as a marker for sperm quality.
We analyzed the effects of this genetic variation on sperm quality traits by using SPSS18.0 software (IBM, Armonk, NY, USA). Multiple comparisons were performed using Tukey’s test. It was expected that bulls with the TCTC genotype would show higher fresh sperm motility and sperm density, and a lower deformity rate than those with the ATCT genotype (p < 0.05). There were no significant differences in ejaculate volume, and post-thaw cryopreserved sperm motility between the genotypes (p > 0.05) (Table 1), indicating that bulls with the TCTC genotype are more likely to show high sperm quality.
Numbers of miRNAs were identified different expressing between high- and low-motility sperms. Bta-miR-6531 is special low expressing in low-motility sperms, the potential molecular function of miR-6531 in sperm has not been elucidated. It was observed that predicted genes of bta-miR-6531 mainly target the calcium and the cAMP signaling pathway. The influence of calcium and cAMP on sperm motility and the relationship between semen parameters and cAMP remains controversial [21]. Moreover, in the bovine Leydig cells model, overexpression of bta-miR-6531 resulted in lower [Ca2+]i and higher MMP than that of the control group. miR-6531 is a retentive miRNA in sperm, that may regulate spermiogenesis in bovine testis. The generation of cAMP can stimulate influx through specific calcium channels, contributing to sperm hyperactivation [22]. Sperm hyperactivation is required to penetrate of the zona pellucida, which is critical for fertilization. As in most cell types, calcium is a key regulator of biofunctions. ATP2A2/SERCA2 is present in mammalian sperm and detected in the acrosome and midpiece areas [23]. During spermatogenesis, ATP2A2 protein expression begins at the primary spermatocyte stage and maintains high signaling levels in the round and elongated spermatids [23]. Typically, ATP2A2/SERCA2 regulate intracellular calcium into endoplasmic reticulum in cells [24]. The level of ATP2A2/SERCA2 of sperm may related the calcium flex. Intracellular calcium is a key regulator of the sperm physiology, including sperm maturation, sperm motility, capacitation and acrosome reaction [25,26]. In present study, we have identified ATP2A2 as a functional target gene of miR-6531 by targeting its CDS region in bovine Leydig cell. ATP2A2 is present in the head and midpiece of the sperm [23], but the functional role of ATP2A2 in sperm are not exactly clear. In this study, we verified a possible relationship between miR-6531 and ATP2A2 in a somatic cell model. We also constructed a linkage between miR-6531, ATP2A2 and calcium in bovine sperm. But this relationship whether occurs accurately in the sperm is still unclear, and the sperm-borne miRNAs might show a little contribution to sperm motility, we just provided a probable event between miR-6531 and ATP2A2. Many extensive researches of expression changes of sperm-borne miRNAs related to sperm quality [27]. Studies on sperm-borne miRNAs have mostly focused on the subsequent fertilization and embryonic development [3,28,29]; thus, the miRNA biological functions in sperm need to be explored in more studies. Genomic SNPs associated with sperm quality have been reported in cattle [8,30] Several functional SNPs in the bta-miR-6531 precursor that are related to sperm quality and its possible molecular mechanism in sperm have been identified. In general, two types of functional SNPs occur in miRNA. Firstly, mutated SNP in target sites might lose the seed region, which is required for binding to target genes [31]. Secondly, SNP may influence the structure of the pre-miRNA, leading to changes in the biogenesis of mature miRNA, effecting the miRNA expression level [9]. The identified variations in pre-bta-miR-6531 in the present study is not located within the target site and did not considerably affect the expression of mature bta-miR-6531 in the sperm with different genotypes. However, bulls with the TCTC genotype showed high fresh sperm motility and density, and a lower deformity rate than that of ATCT genotype. This might indicate that these SNPs are likely associated with better sperm quality. In conclusion, this study revealed miR-6531 regulate calcium signaling by targeting ATP2A2 in bovine Leydig cells and identified a series of SNPs in the precursor of bta-miR-6531 in Holstein bulls, that can be used as molecular biomarkers associated with sperm quality. Also provides an opportunity for a more detailed investigation of the contribution of miRNA-associated SNPs to sperm quality. | true | true | true |
|
PMC9601988 | Valeria Lucci,Elena De Marino,Daniela Tagliaferri,Stefano Amente,Alessandra Pollice,Viola Calabrò,Maria Vivo,Geppino Falco,Tiziana Angrisano | Identification of Cdk8 and Cdkn2d as New Prame-Target Genes in 2C-like Embryonic Stem Cells | 27-09-2022 | PRAME,embryo stem cell,RA-resistant | Embryonic stem cells (ESCs) present a characteristic pluripotency heterogeneity correspondent to specific metastates. We recently demonstrated that retinoic acid (RA) induces an increase in a specific 2C-like metastate marked by target genes specific to the two-cell embryo stage in preimplantation. Prame (Preferentially expressed antigen in melanoma) is one of the principal actors of the pluripotency stage with a specific role in RA responsiveness. Additionally, PRAME is overexpressed in a variety of cancers, but its molecular functions are poorly understood. To further investigate Prame’s downstream targets, we used a chromatin immunoprecipitation sequencing (ChIP-seq) assay in RA-enriched 2C-like metastates and identified two specific target genes, Cdk8 and Cdkn2d, bound by Prame. These two targets, involved in cancer dedifferentiation and pluripotency, have been further validated in RA-resistant ESCs. Here, we observed for the first time that Prame controls the Cdk8 and Cdkn2d genes in ESCs after RA treatment, shedding light on the regulatory network behind the establishment of naïve pluripotency. | Identification of Cdk8 and Cdkn2d as New Prame-Target Genes in 2C-like Embryonic Stem Cells
Embryonic stem cells (ESCs) present a characteristic pluripotency heterogeneity correspondent to specific metastates. We recently demonstrated that retinoic acid (RA) induces an increase in a specific 2C-like metastate marked by target genes specific to the two-cell embryo stage in preimplantation. Prame (Preferentially expressed antigen in melanoma) is one of the principal actors of the pluripotency stage with a specific role in RA responsiveness. Additionally, PRAME is overexpressed in a variety of cancers, but its molecular functions are poorly understood. To further investigate Prame’s downstream targets, we used a chromatin immunoprecipitation sequencing (ChIP-seq) assay in RA-enriched 2C-like metastates and identified two specific target genes, Cdk8 and Cdkn2d, bound by Prame. These two targets, involved in cancer dedifferentiation and pluripotency, have been further validated in RA-resistant ESCs. Here, we observed for the first time that Prame controls the Cdk8 and Cdkn2d genes in ESCs after RA treatment, shedding light on the regulatory network behind the establishment of naïve pluripotency.
Prame, Gm12794c in mice, is a member of a multigene family present in humans and other mammals [1]; it is expressed in different cancers [2] and is associated with a novel stem cell molecular signature in the 2C-like metastate, characteristic of naïve pluripotency [3]. ESCs are characterized by different metastable subpopulations that fluctuate among different levels of pluripotency, defined as metastates. These metastates are marked by specific factors whose expression affects the state of cell pluripotency [4,5,6]. Occasionally, a metastable subpopulation may convert to a highly pluripotent metastate (2C-like) resembling the two-cell stage (2C) embryos, in which a rearrangement of chromatin induces a reprogramming of a transcription named zygotic genome activation (ZGA) [7]. Recently, it was demonstrated that RA enhances ESCs’ metastate subpopulation, named Zscan4 metastate, marked by a specific 2C-like gene signature, including Prame and Zscan4, in which ESCs retain both self-renewal and pluripotency capability [8,9,10]. Correlated to this aspect, human PRAME overexpression has been demonstrated to block retinoic acid (RA)-mediated cell differentiation, cell growth arrest, and apoptotic death, suggesting that PRAME acts as an inhibitor of the retinoic acid receptor (RAR) [11,12,13]. Moreover, PRAME overexpression was found to promote leukemic cell proliferation and inhibit all-trans retinoic acid (ATRA)-induced myeloid differentiation [14] and in several hematological malignancies characterized by the block of myeloid differentiation [15]. Indeed, the murine Prame counteracts RA-dependent differentiation. RA induces high levels of Prame, contributing to the overall DNA hypomethylation and global increase in H3K27 acetylation levels throughout the Cdkn1a transcriptional repression induced by the PCR2 complex [13]. The considerable heterogeneity among the different Prame isotypes made it challenging to determine their specific function, so the clinical relevance of Prame remains still unclear and prompted us to investigate whether Prame is a suitable ATRA-resistance therapeutic target. To better understand this aspect, we performed a Prame chromatin immunoprecipitation followed by sequencing (ChIP-seq). We found Prame-specific binding to Cdk8 and Cdkn2d regulatory regions by this approach. We focused our attention on these two targets because of their involvement in cancer dedifferentiation and pluripotency. Cdk8 is a cyclin-dependent kinase implicated in cellular homeostasis and developmental programming, and it has been shown to regulate several signaling pathways that are crucial regulators of both embryonic stem cell pluripotency and cancer [16]. On the other hand, Cdkn2d is a member of the INK4 family of cyclin-dependent kinase inhibitors that generally regulate the G1-to-S phase transition [17]; interestingly, Cdkn2d dissymmetric mRNA distribution was observed in the 2-cell stage in blastomeres [18]. Our ChIP-seq data were validated by chromatin immunoprecipitation and gene expression analysis in ESpZscan4-EME/Prame-FLAG transgenic cell line. Moreover, we used a second transgenic cell line, overexpressing Prame, to analyze the Prame -specific activity on Cdk8 and Cdkn2d gene expression to maintain the pluripotent metastate in cells treated or not treated with RA.
The pZscan4-Emerald cells, a gift from Dr. Minoru S.H.Ko, were transfected with the Prame-3xFlag vector (ESpZscan4-EME/Prame-FLAG), generated in [13], and validated for pluripotency expression markers after RA treatment. It was necessary to use the FLAG-tag for Prame immunoprecipitation analysis as there is no available antibody against the murine isoform of the Prame studied. ESpZscan4-EME/Prame-FLAG cells were cultured in gelatin-coated 6-well plates in a complete ES medium: DMEM (Sigma-Aldrich, St. Louis, MO, USA); 15% FBS (EuroClone, Italy); 1000 U/mL leukemia inhibitory factor (LIF) (ESGRO, Chemicon, SIGMA, Darmstad, Germany); 1 mM sodium pyruvate; 0.1 mM nonessential amino acids (NEAA), 2.0 mM l-glutamine (Invitrogen, Waltham, MA, USA), 0.1 mM β-mercaptoethanol, and 500 U/mL penicillin/streptomycin, 125 μg/mL G418, and 2.5 μg/mL Blasticidin (Sigma-Aldrich). ESCs were incubated at 37 °C with 5% CO2. To enrich Zscan4-EME positive cells (EME+) expressing Emerald protein, the ESpZscan4-EME/PRAME-FLAG cells were treated with 1.5 μM RA for 4 days. EME+/NoFLAG, containing only Zscan4 promoter fuser to Emerald protein, was used as a negative control in the immunoprecipitation assay with ant-FLAG antibodies. The second ES cell line Prame inducible by tetracycline, ESGm12794cp2Lox was initially described and validated in [13], and in this study it is named ESPramep2Lox. Briefly, the coding sequence of Prame was amplified from an available plasmid and cloned into a p2Lox targeting vector. Cellular clones were grown with 275 μg/mL neomycin (Invitrogen), and an empty p2Lox vector stably transfected in ESCs was used as a negative control cell line. ESPramep2Lox and p2Lox empty vector cell lines were cultured in DMEM (Invitrogen) supplemented with 15% ES-certified FBS (Invitrogen), 0.1 mM nonessential amino acids (Invitrogen), 1 mM sodium pyruvate (Invitrogen), 0.1 mM β-mercaptoethanol (Sigma-Aldrich), 50 U mL−1 penicillin/50 μg mL−1 streptomycin (Invitrogen), and 1000 U mL−1 LIF (Sigma-Aldrich). Clones were treated for 72 h in the absence or presence of 1.5 μg/mL doxycycline to induce transgene expression, followed by the presence or absence of RA 1.5 μM for 4 days. Finally, an E14 ES transient knockdown of the Prame gene was generated. The Prame shRNA was transfected into a pLKO.1 vector (Addgene, Arsenal, UK) using AgeI and EcoRI restriction enzymes. All these passages were verified by sequence analysis. The ES cells were transfected with Lipofectamine Transfection Reagent (InvitrogenTM) according to the manufacturer’s instructions.
1 × 106 cell line ESpZscan4-EME/Prame-FLAG was plated in p100 and treated with 1.5 μM all-trans RA for 4 days and the medium was changed every day. ES E14 cell lines, plated in ES medium and treated with RA, were used as a negative control. Then, the cells were harvested by Trypsin (Invitrogen) and resuspended in a complete ES medium containing 25 mM HEPES buffer and by FACS-sorted according to the fluorescent intensity of Emerald into a complete ES medium containing HEPES and separated in EME+ (Zscan4/Emerald expressing) and EME-(Zscan4/Emerald not expressing) as performed in [19]. Cell sorting experiments were performed by the cell sorter FACSAria and analyzed through FACSDiva Software Version 6.1.3 (Becton Dickinson, Franklin Lakes, NJ, USA). Analysis of forward scatter (FSC) vs. side scatter (SSC) dot plots excluded dead cells and debris. Afterward, an FSC-Area vs. FSC-Height dot plot was used to identify single cells and exclude doublets. In all experiments, purity was higher than 95% of the desired cells.
Total proteins extracted from EME+ and EME− cells were analyzed by western blot as described in [20]. The antibodies used are as follows: mouse monoclonal antibody anti-FLAG M2 (Sigma Aldrich, Milan, Italy) and rabbit polyclonal anti-Gapdh (Santa Cruz Biotechnology, Heidelberg, Germany).
For qPCR analysis of FACS-sorted cells, total RNA was collected immediately after sorting from EME+ and EME- cells by TRIzol (Invitrogen) as performed in [21]. The integrity of the RNA was assessed by denaturing agarose gel electrophoresis (presence of sharp 28S, 18S, and 5S bands) and spectrophotometry. A total of 1 μg of RNA of each sample was reverse transcribed with QuantiTect® Reverse Transcription (Qiagen, Germany, Europe) according to the manufacturer’s instructions. qPCR analyses were performed using 20 ng cDNA per well in duplicate with SYBR green master mix (Applied Biosystems, Waltham, MA, USA) according to the manufacturer’s instructions. Reactions were run on Applied Biosystem 7500 (Applied Biosystems). Fold induction was calculated and normalized with the DDCt method and considered the values of at least three independent experiments. The gene-specific primers are available in Table 1.
ChIP analysis was performed as previously described [22] using the EpiQuikTM chromatin immunoprecipitation kit from Epigentek Group Inc. (Brooklyn, NY, USA). Then, 2 × 106 cells sorted into EME+ and EME- were cross-linked in a fresh culture medium with 1% formaldehyde for 10 min, sonicated and used for each immunoprecipitation with 3 μg of antibodies. Immunoprecipitated DNA was used for massively parallel sequencing or analyzed on a real-time PCR machine (Applied Biosystem 7500), SYBR green master mix according to the manufacturer’s instructions. Primer sequences are listed in Table 1.
Prame immunoprecipitation was used as anti-FLAG antibody on cells EME+ and EME- using independent biological replicates. Single-end 50-bp libraries were prepared from Genomix4life. The signal obtained by precipitation with the control EME- and input was subtracted from the signals obtained with the specific antibodies. Results are expressed as a percentage of the input, and calculations considered the values of at least two biological replicates. ChIP-seq libraries were prepared with 10 ng of ChIP (or Input) DNA with TruSeq ChIP Sample Prep Kit according to the manufacturer’s instructions of Illumina sequencing. Before sequencing, libraries were quantified using Qubit (Invitrogen) and quality-controlled using Agilent’s bioanalyzer. The basic steps are represented in the form of a flowchart in Figure 1. A 50 bp single-end sequencing was performed using Illumina HiSeq 2000 platform (Genomix4life S.R.L., Baronissi, Salerno, Italy) according to standard operating procedures. ChIP-seq reads were quality-checked with NGS QC Toolkit. Alignments were performed with BWA to the reference genome mm9 (mouse assembly July 2007 NCBI37). SAMtools [23] and BEDtools [24] were used for filtering steps and file format conversion. The peaks were identified from uniquely mapped reads without duplicates using MACS, and the p-value cutoff used for peak detection was 1 × 10−5. ChIP-seq data were normalized on DNA Input and expressed as a log2 value. ChIP-seq peaks were annotated with PAVIS [25]. IGV genome browser was used for data visualization (Figure 1). To plot data of average profiles around prime target genes, specific site positions were retrieved from the mouse genome (mm9).
Statistical significance between two or three groups was assessed with Student’s t-test and or ANOVA test, respectively. Data are expressed as mean ± standard deviation (SD). All experiments were repeated at least three times. A p-value < 0.05 was considered statistically significant.
To identify the Prame target genes, we used a transgenic ESCs line that simultaneously expresses FLAG-tagged Prame protein and Emerald downstream to the Zscan4 promoter (ESpZscan4-EME/Prame-FLAG), as described in Material Methods and previously validated for expression of pluripotent marker genes [13]. ESpZscan4-EME/Prame-FLAG cell line RA-treated was sorted for Emerald expression (or Zscan4 metastate), then immunoprecipitated with an anti-FLAG antibody and analyzed by next-generation sequencing (Figure 2A). To obtain an enriched EME+ cell population, ESpZscan4- EME/Prame-FLAG cells were cultured with 1.5 μM of RA for four days. Zscan4 induction was evaluated through fluorescence microscopy (Figure 2B) and FACS analysis to quantify Zscan4 expression; we analyzed the percentage of green cells (EME+) in ESpZscan4- EME/Prame-FLAG that increased to 21.8% after RA treatment (Figure 2C). In addition, we further confirmed the successful separation of Prame gene and protein expression in EME+ cells by western blot (Figure 2C). Finally, we performed a ChIP-seq analysis in EME+ cells.
To identify Prame target genes upon RA treatment, we collected 2 × 106 EME+ cells after sorting. In Figure S1A, we can see the distribution of Prame binding regions: upstream and downstream gene regions, 5′UTR and 3′UTR, introns in the form of a pie chart (Figure S1B). As shown in Figure S1B, binding is mainly found in regulatory regions (upstream, 3′-UTR and downstream gene regions). The ChIP-seq analysis showed 15 specific genes significantly bound by Prame (Cdk8, Cdkn2d, uc009gks.1, St6galnac1, Stard13, Mir26b, 533042B09Rik, Gm15319, Gm3168, AK053193, Mir715, Zc3h7a, Filip1l, AK041614, and Gm10406) and described in Table 2. Among these, we chose Cdk8 and Cdkn2d for further validation. Cdk8 is a cyclin-dependent kinase that maintains both tumors and embryonic stem cells in an undifferentiated state [16], while Cdkn2d (cyclin-dependent kinase inhibitor 2D) is a cell growth regulator that controls G1 cell cycle progression and results in dissymmetrically expressed cells in the 2-cell stage, thus, representing good targets of the pluripotency stage marked by Prame expression [14]. Figure 3A indicates the schematic position of primers used for ChIP analysis validation by qPCR on Cdk8 and Cdkn2d enriched sequences after ChIP-seq. EME+ and EME- cells and EME+/NoFLAG DNA immunoprecipitated for Prame, with anti-FLAG antibody, was analyzed through qPCR for Cdk8 and Cdkn2d enrichment to confirm data obtained by ChIP-seq. Data are calculated as the percentage of the input (Figure 3B). The EME+/NoFLAG cells were used as a negative control of immunoprecipitation.
To understand how RA influences Cdk8 and Cdkn2d expression, a qPCR analysis was performed in EME+ and in EME- cells (Figure 4A,B). The gene expression analysis showed a significant enhancement of Cdk8 and Cdkn2d expression in EME+ cells compared to EME- (Figure 3). These data indicate that the Prame binding to Cdk8 and Cdkn2d genes increases their expression.
To understand whether Prame overexpression influences Cdk8 and Cdkn2d gene expression in ES cells, an ESPramep2Lox cell line in which Prame expression is induced by doxycycline was employed [13]. The ESCs stably transfected with the inducible PRAME expressing construct were treated with doxycycline for 3 days (Figure 5A) and then with RA for 4 days. To validate Prame induction by doxycycline and then by RA treatment, both Prame and Zscan4 gene expression were analyzed by qPCR (Figure 5B). The ES cell line p2Lox empty vector was used as a negative control for the Dox treatment. Our data shows a 5-fold higher Prame gene expression after the Dox treatment (Figure 5B), with cells maintaining a classic ESC pluripotent morphology as previously described [13]. Zscan4 gene expression was observed after RA treatment, confirming the Zscan4 metastate induced by RA. Finally, we observed that Cdk8 and Cdkn2d expressions were directly increased in ESCs overexpressing Prame without RA (Figure 5C). This data shows for the first time that Prame directly influences the expression of their target genes, also without RA treatment.
To shed light on the role of Prame in Cdk8 and Cdkn2d gene regulation, transient inactivation of Prame was performed in E14 ES. The shPrame vector and a sh-scrambled (sh-scr as a negative control) were transfected in ESCs for 24 h. Prame specific-shRNA mapped to the Exon2-coding region (Figure 6A). Total RNA was extracted from ES cells transfected with shPrame and sh-scr. The qPCR analysis showed a significant reduction in Cdk8 expression in ESCs upon Prame silencing compared to sh-scr cell control (Figure 6B), while Cdkn2d expression remains unaffected by Prame silencing (Figure 6B).
The role of PRAME in human cancers is well established. PRAME is recruited in human cells to epigenetically and transcriptionally activate promoter regions bound by the nuclear transcription factor Y (NFY). This transcription factor is essential for early embryonic development [26] and has been implicated in the maintenance of the high proliferative capacity of ES cells [27] as well as in the inhibition of differentiation [28,29]. In addition, several studies indicated that Prame-like genes have roles in the early stages of spermatogenesis [30] and oogenesis [31], as well as in embryonic development and embryonic stem cells [32,33]. Moreover, Prame overexpression in E14 embryonic stem cells was reported to maintain a pluripotent state in the absence of the antidifferentiation factor LIF [13,34] and in ESCs that were RA-resistant [8]. Together, all these data are consistent with the notion that stem cells and neoplastic tissues share many properties, and several oncogenic pathways can also regulate self-renewal mechanisms in stem cells [35]. Therefore, the identification of Prame functions and targets indeed represents a step forward, not only in the cancer diagnostic approach but also in the therapeutic approach. Here, we report, for the first time, two putative novel Prame targets in RA-treated ESCs. To better understand the regulatory network underpinning the establishment of naïve pluripotency marked by Zscan4 and Prame genes in ESCs, we used ESpZscan4-EME/Prame-FLAG transgenic cell lines, allowing us to immunoprecipitate Prame and its chromatin-bound fragments by using anti-FLAG antibodies. The Prame-bound sequences were characterized by ChIP-seq analysis. Among the putative Prame targets identified, we focused our attention on Cdk8, and Cdkn2d. Cyclin-dependent kinases (CDKs) are key players in cell cycle regulation. Being involved in the regulation of cell cycle checkpoints, Cdk8, and Cdkn2d are involved in several cellular pathways, however, very little is known about their involvement in the establishment of RA-resistant Zscan4 metastate pluripotent cells. CDK8 was shown to maintain tumor dedifferentiation and embryonic stem cells pluripotency stage [16] drugs resistance in several cancer disease models [36,37]. Similar evidence was reported for Cdkn2d overexpression [38]. Concerning Cdkn2d, unlike Cdk8, we could not find any alteration in its expression upon Prame silencing. A possible explanation for this phenomenon is that Cdkn2d may be regulated by a higher expression level of Prame occurring under particular cell conditions. Our data uncovered an important novel role of Cdk8 and Cdkn2d expression in the Prame-dependent naïve pluripotent metastate. Remarkably, CDK inhibitors, such as CDK4/6 inhibitors, are already used in preclinical studies for cancer treatment [39]. In particular, CDK8 inhibitors have been used in acute myelogenous leukaemia and several other types of cancers, including breast cancer. Cancer cells treated with CDK8 inhibitors responded with decreased cell viability and increased apoptosis [40,41]. However, the function of Cdk8 and Cdkn2d may be cell-context dependent. Therefore, it will be necessary to deeply investigate the protein expression and the molecular and epigenetic mechanisms underlying Prame-induced Cdk8 and Cdkn2d activation in ESCs to identify the tumor specificity of anti-Cdk8 or anti-Cdkn2d pharmacological compounds affecting cancer cells mimicking the PRAME-marked stage of cell pluripotency.
In our work, we identified, for the first time, Cdk8 and Cdkn2d as new Prame -target genes through a ChIP-seq analysis specific to ESCs enriched in Prame expression after RA treatment. The identification of new targets for Prame in ESCs represents a milestone in the field of ESC therapy, specifically for CSC drug delivery. Our data opens a new window on the study of CSC therapies, RA-resistant. In the future, it will be necessary to use drugs and combinations of these in cancers marked by Cdk8 and Cdkn2d expression. | true | true | true |
|
PMC9602136 | Ambreen Iqbal,Haibin Yu,Ping Jiang,Zhihui Zhao | Deciphering the Key Regulatory Roles of KLF6 and Bta-miR-148a on Milk Fat Metabolism in Bovine Mammary Epithelial Cells | 09-10-2022 | bovine mammary epithelial cells,bta-miR-148a,KLF6,milk fat | MicroRNAs (miRNAs) are non-coding RNAs that regulate the expression of their target genes involved in many cellular functions at the post-transcriptional level. Previously, bta-miR-148a showed significantly high expression in bovine mammary epithelial cells (BMECs) of Chinese Holstein cows producing high milk fat compared to those with low milk fat content. Here, we investigated the role of bta-miR-148a through targeting Krüppel-like factor 6 (KLF6) and further analyzed the role of KLF6 in regulating fat metabolism through targeting PPARA, AMPK/mTOR/PPARG, and other fat marker genes in BMECs of Chinese Holstein. The bioinformatics analysis showed that the 3’ UTR of KLF6 mRNA possesses the binding sites for bta-miR-148a, which was further verified through dual-luciferase reporter assay. The BMECs were transfected with bta-miR-148a-mimic, inhibitor, and shNC, and the expression of KLF6 was found to be negatively regulated by bta-miR-148a. Moreover, the contents of triglyceride (TG), and cholesterol (CHO) in BMECs transfected with bta-miR-148a-mimic were significantly lower than the contents in BMECs transfected with bta-miR-148a-shNC. Meanwhile, the TG and CHO contents were significantly increased in BMECs transfected with bta-miR-148a-inhibitor than in BMECs transfected with bta-miR-148a-shNC. In addition, the TG and CHO contents were significantly decreased in BMECs upon the down-regulation of KLF6 through transfection with pb7sk-KLF6-siRNA1 compared to the control group. Contrarily, when KLF6 was overexpressed in BMECs through transfection with pBI-CMV3-KLF6, the TG and CHO contents were significantly increased compared to the control group. Whereas, the qPCR and Western blot evaluation of PPARA, AMPK/mTOR/PPARG, and other fat marker genes revealed that all of the genes were considerably down-regulated in the KLF6-KO-BMECs compared to the normal BMECs. Taking advantage of deploying new molecular markers and regulators for increasing the production of better-quality milk with tailored fat contents would be the hallmark in dairy sector. Hence, bta-miR-148a and KLF6 are potential candidates for increased milk synthesis and the production of valuable milk components in dairy cattle through marker-assisted selection in molecular breeding. Furthermore, this study hints at the extrapolation of a myriad of functions of other KLF family members in milk fat synthesis. | Deciphering the Key Regulatory Roles of KLF6 and Bta-miR-148a on Milk Fat Metabolism in Bovine Mammary Epithelial Cells
MicroRNAs (miRNAs) are non-coding RNAs that regulate the expression of their target genes involved in many cellular functions at the post-transcriptional level. Previously, bta-miR-148a showed significantly high expression in bovine mammary epithelial cells (BMECs) of Chinese Holstein cows producing high milk fat compared to those with low milk fat content. Here, we investigated the role of bta-miR-148a through targeting Krüppel-like factor 6 (KLF6) and further analyzed the role of KLF6 in regulating fat metabolism through targeting PPARA, AMPK/mTOR/PPARG, and other fat marker genes in BMECs of Chinese Holstein. The bioinformatics analysis showed that the 3’ UTR of KLF6 mRNA possesses the binding sites for bta-miR-148a, which was further verified through dual-luciferase reporter assay. The BMECs were transfected with bta-miR-148a-mimic, inhibitor, and shNC, and the expression of KLF6 was found to be negatively regulated by bta-miR-148a. Moreover, the contents of triglyceride (TG), and cholesterol (CHO) in BMECs transfected with bta-miR-148a-mimic were significantly lower than the contents in BMECs transfected with bta-miR-148a-shNC. Meanwhile, the TG and CHO contents were significantly increased in BMECs transfected with bta-miR-148a-inhibitor than in BMECs transfected with bta-miR-148a-shNC. In addition, the TG and CHO contents were significantly decreased in BMECs upon the down-regulation of KLF6 through transfection with pb7sk-KLF6-siRNA1 compared to the control group. Contrarily, when KLF6 was overexpressed in BMECs through transfection with pBI-CMV3-KLF6, the TG and CHO contents were significantly increased compared to the control group. Whereas, the qPCR and Western blot evaluation of PPARA, AMPK/mTOR/PPARG, and other fat marker genes revealed that all of the genes were considerably down-regulated in the KLF6-KO-BMECs compared to the normal BMECs. Taking advantage of deploying new molecular markers and regulators for increasing the production of better-quality milk with tailored fat contents would be the hallmark in dairy sector. Hence, bta-miR-148a and KLF6 are potential candidates for increased milk synthesis and the production of valuable milk components in dairy cattle through marker-assisted selection in molecular breeding. Furthermore, this study hints at the extrapolation of a myriad of functions of other KLF family members in milk fat synthesis.
Milk is a nutrient-rich white liquid obtained from mammals’ mammary glands, and it is a chief nutrient source for infant mammals [1]. Milk fat is a vital source of energy owing to the nutrients such as bioactive lipids and fat-soluble vitamins [2]. In milk fat, the principal component is triacylglycerol (TAG), which comprises about 97–98%. In contrast, the other minor components include diacylglycerol (DAG), which comprises approximately 0.28 to 0.59 percent of weight, and monoacylglycerol (MAG) has about 0.16 to 0.38 wt percent range. The free fatty acids are just about 0.1 to 0.44 wt %, while the CHO component is about 0.4 to 0.45%. [3]. MicroRNAs are one of the molecular players in regulating milk fat synthesis in mammary epithelial cells. MiRNAs belong to the family of non-coding RNAs that is about 22 nucleotides long. They regulate the expression of their target genes by binding to the 3’ untranslated region (3’ UTR) of their mRNAs [4]. The expression of miRNAs in the mammary gland arbitrates important functions, including mammary gland development, fat metabolism, and lactation [5,6]. Several researchers have investigated the roles of miRNAs in dairy lactation or mammary gland development. Recently, Do et al. found that 58 miRNAs were dynamically and differently expressed across lactation phases and that 19 miRNAs were significantly and time-dependently expressed during lactation [7]. Furthermore, lipid metabolism-related miRNAs’ role in lipid metabolism has been reported [8]. Several lipid metabolism-related miRNAs including miR-33 [9], miR-122 [10], miR-370 [11], miR-378/378 * [12], miR-143 [13], miR-335 [14], and miR-103 [15] have been reported. Moreover, miR-103 [15], miR-2885, miR-135a, and miRNA-370 are involved in the metabolism of glucose and lipids [16]. The miR-148a belongs to a highly conserved family, the miR-148/miR-152 family, and is abundantly expressed in mammary gland tissue [17,18]. Previous studies reported the role of miR-148a in the differentiation of the C2C12 myoblast and skeletal muscle cells and myoblast differentiation into the myotube [19]. Van Wijnen et al. [20] stated that miR-148a was implicated in osteoclast development and caused the monocyte to osteoclast transition. Gailhouste et al. [21] indicated that improved miR-148a expression could induce liver cell differentiation and maturation by the inhibition of DNA (cytosine-5-)-methyltransferase 1 (DNMT1). The high expression of miR-148a in the adult liver controls cholesterol and triglyceride homeostasis [22]. In addition, miR-148a could facilitate the differentiation of primary adipocytes to mature adipocytes [23]. Additionally, adipocyte differentiation and regeneration are both affected by miR-148a and miR-17-5p [24]. Remarkably, miR-148a and miR-17-5p seemed to increase the triglyceride content and lipid accumulation of goat MECs [25]. In addition, miR-148a also shows high milk expression as a component of milk exosomes and milk fat globules [26]. Intriguingly, single-nucleotide polymorphisms (SNPs) in the promoter region of miR-148a have recently been linked to changing LDL-C and triglyceride levels in humans, implying a putative physiological role for this miRNA in regulating lipid metabolism and thus emphasizing it for further research [27,28,29]. Bta-miR-148a also showed significantly high expression in BMECs from high-fat Chinese Holstein BMECs compared to the low-fat Chinese Holstein BMECs [30]. Krüppel-like factors (KLFs), a family of transcription factors (TFs), bind preferably to the GC-rich sequence CACCC through its three conserved zinc finger motifs toward the C-terminal side of 23–25 amino acids [31]. KLF6 binds to different acetyl-transferases, including cyclic adenosine monophosphate response element-binding protein (CREBBP), p300, and p300/CBP-associated factor [32]. The binding of KLF6 to these acetyl-transferases leads to the acetylation of KLF6, which stimulates its transcriptional activity [33]. It can also lead to the acetylation of histones turning into chromatin remodeling and transcription initiation in areas targeted by KLF6 [33]. Furthermore, the nuclear receptor peroxisome proliferator-activated receptor α (PPARA) regulates the expression of genes related to FA oxidation, lipid uptake, lipid transport, and β-oxidation [34]. Through their engagement with PPARA, three members of the KLF family enhance FA oxidation in the heart [34]. The PPARs belong to the nuclear receptor family and ligand-inducible transcription factors [35]. The family of PPARs is made up of three members: PPARA (also known as NR1C1), PPARB (also known as NR1C2), and PPARG (also known as NR1C3). PPARs are combined with retinoid X receptors to form heterodimers. They govern the network gene expression engaged in several processes such as lipid metabolism, adipogenesis, metabolic balance, and inflammation [36]. The first PPAR to be recognized as PPARA is located in the liver and heart. Hypolipidemic fibrate medications decrease cholesterol levels by targeting this enzyme, which regulates fatty acid oxidation [37]. PPARs are also present in brown adipose kind of fat. Note that whereas PPARG (also known as PPARB and more generally referred to by its abbreviation PPARG/B) shares many of the same activities as its counterpart, it is articulated in a wider range of tissues, including the heart, skeletal muscle, and liver [36,37]. PPARG is a master controller of adipogenesis and an effective modulator of whole-body insulin sensitivity and lipid metabolism [35,38]. Due to an alternative promoter and splicing use, PPARG has two isoforms, PPARG1, and PPARG2, at its N terminus having an extra 30 amino acids [35,38,39]. High fat diet can stimulate the expression of PPARG2 in various organs, although PPARG1 is expressed in numerous tissues [39,40]. In vitro cultured bovine mammary epithelial cells (BMECs) are widely used as an experimental cell model for studying milk production or its components due to their milk synthesizing ability. Previously, our group reported the role of different miRNAs in regulating the expression of milk-related genes in BMECs. Therefore, BMECs were used in this study as the experiment object to verify the relationship between bta-miR-148a and the target gene KLF6. The role of bta-miR-148a on the content of triglyceride, cholesterol, and free fatty acids in BMECs after the overexpression or inhibition of bta-miR-148a and by the construction of pBI-CMV3-KLF6, pBI-GFP-Neo-CMV3, pb7sk-KLF6-siRNA, and pb7sk-GFP-Neo was also investigated. To the best of our knowledge, this is the second study to report the role of bta-miR-148a in regulating fat metabolism. In addition, it is the first study in BEMCs that elucidates the role of KLF6 in regulating fat metabolism through targeting PPARA, AMPK/mTOR/PPARG, and other fat marker genes in BMECs of Chinese Holstein.
The BMECs of Chinese Holstein dairy cows were isolated and maintained in the animal genetics and breeding laboratory, college of Animal Science, Jilin University [41] by following the guidelines for the care and use of laboratory animals of Jilin University (Animal Care and Use Committee permit number: SY201901007) [42,43,44].
The plasmids used in this study include bta-miR-148a mimics, inhibitors, and negative controls synthesized by (Shanghai Gene Pharma Corporation Technology Co., Ltd., Shanghai, China) and cloned in E. coli DH5α cells (Vazyme, C502-03 Nanjing, China). The plasmid was extracted using (Endo free Maxi Plasmid kit, Tiangen Biotech Co., Ltd., DP117, Beijing, China). Transfection was performed using Fugene HD Transfection Reagent (Promega Biotechnology Co., Ltd., Beijing), Total RNA Extraction Reagent TRIzol (15596-026, Invitrogen, Waltham, MA, USA), a Reverse Transcription Kit (Vazyme, Hiscript II qRT SuperMix for qPCR (gDNA wiper) R223-01 China), and qPCR (Vazyme, ChamQ Universal SYBR qPCR Master Mix, Q711, Nanjing, China). Primers were synthesized by Genewiz (Suzhou, China). Chloroform, absolute ethanol, and isopropanol of the domestic analytical pure category were from Tianjin Concord Technology. Xho I restriction enzyme (K2704AA) was purchased from TaKaRa, Not I restriction enzyme (0551712) was purchased from Biolabs, and T4 DNA ligase (00758655) was purchased from Thermo Scientific. PBS was obtained from Sangon Biotech (GC10FA0002, Shanghai, China). Protein was extracted using RIPA lysis buffer (Dalian Meilun Biological Technology Co. Ltd., Bejing, China, PN: MA0171). Protein concentration was determined with a TaKaRa BCA Protein Assay Kit, (T9300A) and KLF6 antibody (bs-1395R, Bioss Co. Ltd.). A dual-luciferase reporter assay kit (Vazyme, cat# DL101-01, Nanjing, China), triglyceride assay kit (Applygen, Beijing, China), and cholesterol assay kit (Applygen, Beijing, China) was used during the whole research work.
The following experimental apparatus and instruments were used in this study: 5% CO2 cell culture incubator (Thermo, Marietta, OH, USA) and a fluorescent microscope (TE2000, Nikon, Tokyo, Japan). cDNA was prepared on a Bio-Rad T100 thermal cycler, qPCR was performed using the Bio-Rad CFX connect real-time system (USA), and the quality of RNA was measured using a spectrophotometer (UNIC2802H, Shanghai, China). The Wet/Tank Blotting Systems (Bio-rad, Hercules, CA, USA) were used to perform the Western blotting. The signal intensities were captured by a Tanon 5200 chemiluminescence/fluorescence image analysis system (Tanon, Urumqi, China). The luciferase activity and TG contents were detected by the SpectraMax M5 microplate reader (Molecular Devices, San Jose, CA, USA). Other apparatus involved in this study included a liquid nitrogen tank (Chart/Golden Phoenix Liquid Nitrogen Container USA), Microplate Centrifuge (Tiangen Biotech Co., Ltd., Beijing, China), and Gel Imager (Alpha Innotech, San Leandro, CA, USA).
The mature sequence of bta-miR-148a was identified from (http://www.mirbase.org (accessed on 6 February 2021)), and the binding sites for bta-miR-148a in 3’ UTR of KLF6 were found through TargetScan (http://www.targetscan.org (accessed on 10 February 2021)). The primers for qPCR (Table 1) and luciferase activity (Table 2) used in this study were designed by Primer Premier 6.0 (Premier Biosoft International, Canada). Each experimental step/reaction was repeated three times, and GAPDH or U6 were used as a reference where needed. Primers (Table 1) were designed for the 3’ UTR, and PCR amplified the 3’ UTR of KLF6.
The bioinformatic analysis predicted that KLF6 has two bta-miR-148a binding sites in its 3’ UTR. PCR amplified the 3’ UTR sequence of KLF6 for Not I and Xho I restriction fragments. Then, the amplified sequence was cloned into a pmiR-RB-REPORTTM vector to construct the vector of pmiR-RB-REPORTTM-KLF6-WT/MUT. The sequences for the DNA oligos used in this particular assay are listed in Table 2. These vectors of pmiR-BR-REPORTTM-KLF6-WT/MUT were co-transfected with bta-miR-148a-mimic into the BMECs using the same transfection conditions mentioned above except that the used concentration of pmiR-BR-REPORTTM-KLF6-WT/MUT and bta-miR-148a-mimic was 1 μg. The Dual-Glo luciferase assay system (Promega, USA) was used to detect the firefly (hluc+) and Renilla (hRluc) luciferase activities according to the manufacturer’s protocols. The hluc+ luciferase was used as the reference to correct the transfection efficiency variation, and the relative luciferase activities were calculated as hRluc/hluc+. A spectroMax M5 microplate reader detected the fluorescence values, and the results were analyzed using GraphPrism8 software. This experiment was conducted on three replicates with the same number of cells and transfected with the same vectors using the same culture conditions. The KLF6 overexpression vector, pBI-CMV3-KLF6, was a kind gift from Lixin Xia of Jilin University. For the KLF6 interference vector construction, two complementary oligos with restriction cutting sites on the 5′ end was designed using the siRNA Construct Builder tool of GenScript Biotech and synthesized by Sangon Biotech (Shanghai; Table 3). For the siRNA cloning, 10 µL of the reaction mixture was made from 4 µL of NEB buffer, 4 µL of ddH2O, 1 µL of F, and 1 µL of R primer. Then, this reaction mixture was put in the T100 Thermal Cycler of Bio-Rad for the annealing purpose using different temperature gradients for specific times including 90 ℃ for 5 min, and 80 ℃, 70 ℃, 60 ℃, 50 ℃, 40 ℃, 30℃, 20 ℃, and 16 ℃ all for 30 s each. For the ligation purpose, 10 µL of the reaction mixture was prepared from 1 µL of T4 ligase buffer, 2 µL of annealed product from the previous step, 5 µL of pb7sk-GFP-Neo, and 2 µL of ligase buffer. We kept this reaction mixture for 5 h at 16 ℃. After 5 h, 50 µL of DH5α cells were taken, and 10 µL of connecting liquid was added and placed in ice for 30 min after gentle mixing. Then, the mixture was placed in the water bath at 42 ℃ for 60–90 s and placed back on the ice for 3 min. Following this step, 450 µL of LB media was added, and the mixture was put for 45 min at 37 ℃. The mixture was then spread uniformly on the solidified LB media in a Petri plate. The Petri plate was placed in an incubator at 37 ℃ for 12–16 h. Following the growth of bacterial colonies, the good colonies were selected and put in LB medium, which was shaken for 5 h at 37 °C. About 500 µL of sample from turbid solution was sent to Sangon Biotech for sequence matching. After sequence matching, the plasmid was extracted using the EndoFree Maxi Plasmid kit (Tiangen Biotech Co., Ltd. Beijing, China) following the supplier’s instructions.
The culture conditions for BMECs included basal media consisting of DMEM/F12 (Hyclone, Grand Island, NY, USA) supplemented with 10% (v/v) fetal bovine serum (Pasching, Austria), and the cells were incubated with 5% CO2 at 37℃. For the transfection of bta miR-148a-mimic, inhibitor, and shNC plasmids, the BMECs were seeded in a six-well cell culture plate (Nest, Wuxi, China) with about 1 × 106 cells/well. The old culture media was removed upon reaching 70–80% cell confluency, BMECs were washed twice with PBS (Sangon Biotech, Shanghai, China), and fresh culture media was added. The transfection system included 2 μg of plasmid, 6 μL of Fugene, and 200 μL of DMEM/F12 for each well. After mixing, the transfection mixture was incubated for 15 min at room temperature, then transferred to the BMECS in a six-well plate, and cells were put back in the incubator. After 24 h, the cells were examined under a fluorescence microscope for cell morphology and evaluation of transfection through the expression of green fluorescent protein (GFP). There was a total of three replicates with the same number of cells and the same culture and transfection conditions. The same protocol was used for the transfection of pb7sk-KLF6-siRNA, pb7sk-GFP-Neo, pBI-CMV3-KLF6, and pBI-GFP-Neo-CMV3 into BMECs.
After 24 h of transfection of BMECs with miR-148a-mimic, miR-148a-inhibitor, and miR-148a-negative control (shNC) vectors, the total RNA was extracted using the TRIzol reagent following the manufacturer’s protocol. The purity and RNA concentration was determined by agarose gel electrophoresis and spectrophotometer (Thermo NANODROP-2000). Then, 1 μg of total RNA was reverse transcribed to cDNA using the Reverse Transcription Kit (VAZYME, Hiscript II qRT SuperMix for qPCR (gDNA wiper). For the KLF6 overexpression, interference, and control group of BMECs after 24 h of respective transfection with pBI-CMV3-KLF6, pBI-GFP-Neo-CMV3, pb7sk-KLF6-siRNA, and pb7sk-GFP-Neo, the same methods were used for RNA extractions, purity and concentration measurement, and cDNA synthesis was used in the bta-miR-148a experiment. Real-time quantitative PCR was performed in a 20 μL reaction mixture including the following components: 10 μL of SYBR Mix, 0.5 μL of each of the upstream and downstream primers, 1 μL of cDNA, and 8 μL of ddH2O. The reaction procedure included the following step; pre-denaturation at 95 ℃ for 30 s, denaturation at 95 ℃ for 5 s, and annealing at 60 ℃ for 30 s with 35 cycles of repetition. The final values were calculated through 2−ΔΔCt by using GAPDH as a reference. GraphPadPrism8 software was used for the statistical analysis, and the unpaired t-test function was used to compare different groups for differential expression analysis. All experiments were conducted in three replicates using the same number of cells and culture and transfection conditions. Real-time PCR primer sequences of PPARG and PPARA pathway related, and the marker genes of lipid synthesis are shown in Table 4.
Total RNA from BMECs was extracted 24 h after transfection through the steps mentioned above RNA extraction. After that, the total RNA concentration was detected, and cDNA was reverse transcribed using the Reverse Transcription Kit (VAZYME, Hiscript II QRT SuperMix for qPCR (gDNA wiper). Using cDNA as a template, the SYBR Premix Ex-Taq TM fluorescence quantitative kit was used for qRT-PCR.
The total protein was extracted from BMECs 24 h post-transfection of bta-miR-148a-mimic, bta-miR-148a-inhibitor, and bta-miR-148a-shNC. Protein was extracted using RIPA lysis buffer (Dalian Meilun biological technology Co. Ltd., PN: MA0171). First, the BMECs were washed twice with PBS, and 200 μL of RIPA/well of a six-well cell culture plate was added, and cells were incubated at 4 ℃ for 30 min. Then, the cell mixture was collected in 1.5 mL centrifuge tubes and centrifuged at 12,000 rpm for 20 min. The supernatant was collected, and protein concentration was determined with a TaKaRa BCA Protein Assay Kit, T9300A). Then, Western blotting was performed by making SDS-PAGE (Epizyme, Biotechnology Co., Ltd. Shanghai, China, PG112) gel electrophoresis, and protein was run on the gel. For immunoblotting, primary polyclonal-KLF6 antibody and monoclonal GAPDH at a 1:500 dilution and a 1:1000 dilution were used, respectively. The secondary antibody used was the HRP-conjugated anti-rabbit antibody with 1:2000 dilution (BioWorld, Irving, TX, USA, BS13271). The signal intensities were captured by a Tanon 5200 chemiluminescence/fluorescence image analysis system (Tanon, China). This experiment was performed on three replicates using the same number of cells and culture and transfection conditions. The same protocol was used for the Western blot of KLF6 protein in pb7sk-KLF6-siRNA1, pb7sk-GFP-Neo, pBI-CMV3-KLF6, and pBI-GFP-Neo-CMV3 transfected BMECs.
The BMECs were transfected with bta-miR-148a-mimic, inhibitor, and shNC in six-well cell culture plates to determine TG contents. After 24 h of successful transfection, the intracellular triglyceride content of BMECs was extracted and detected with the help of a tissue and cell triglyceride assay kit (APPLYGEN, E1013, Beijing, China) following the manufacturer’s protocols. The extracted samples were examined using the software Gen5 CHS (SYNERGY|HTX multi-mode reader, Bio Tek, S1LFTA), and the total TG content was adjusted by the quantity of total protein. Each sample was replicated thrice, and the final results were calculated using the average values. Moreover, this experiment was repeated three times following the same steps mentioned here. The same protocol was used for the intracellular triglyceride content of BMECs transfected with pb7sk-KLF6-siRNA1, pb7sk-GFP-Neo, pBI-CMV3-KLF6, and pBI-GFP-Neo-CMV3.
For the determination of CHO, the BMECS were transfected with bta-miR-148a-mimic, inhibitor, and shNC in six-well cell culture plates. The intracellular cholesterol content of BMECs was extracted and detected by the tissue and cell cholesterol assay kit (APPLYGEN, China, E1015) following the manufacturer’s protocols. The extracted samples were examined using the software Gen5 CHS (SYNERGY|HTX multi-mode reader, Bio Tek, S1LFTA), and the total CHO content was adjusted by the quantity of total protein. Each sample was replicated thrice, and the final results were calculated using the average values. Moreover, this experiment was repeated three times following the same steps mentioned here. The same protocol was used to determine CHO in pb7sk-KLF6-siRNA1, pb7sk-GFP-Neo, pBI-CMV3-KLF6, and pBI-GFP-Neo-CMV3 transfected BMECs.
All results were presented as the means ± standard error of the mean (SEM) of separate experiments (n ≥ 3). The significant differential analysis among different groups was performed using unpaired t-tests in GraphPad Prism8 software (San Diego, CA, USA). Statistical significance is presented as * p <0.05, ** p <0.01, *** p <0.001.
After 24 h of transfection of BMECs with bta-miR-148a mimic, inhibitor and shNC, the BMECs were observed under a fluorescence microscope to observe the cell morphology and transfection efficiency. The morphology of cells remained unchanged, and the obvious expression of GFP verified the successful transfection (Figure 1A–C) and the transfection efficiency of 80–90% in the transfected group was observed. The successful transfection witnessed through GFP was further verified by investigating the relative expression of bta-miR-148a in BMECs transfected with bta-miR-148a mimic bta-miR-148a inhibitor and the shNC group using qPCR. The results about the expression trend of bta-miR-148a suggest that its relative expression in bta-miR-148a-mimic transfected BMECs was significantly higher (p < 0.001) in comparison to that of the shNC group with a 1.4-fold change between bta-miR-148a-mimic and the shNC group (Figure 2). On the other hand, a significantly lower (p < 0.001) relative expression of bta-miR-148a-inhibitor transfected BMECs as compared to that of the shNC group was observed at a 0.6-fold change (Figure 2). These results manifest that the plasmids for bta-miR-148a-mimic, inhibitor and shNC were successfully transfected and showed their respective expression in the BMECs.
Over a dozen target genes linked to lipid metabolism were screened in addition to KLF6 via bioinformatics prediction. From these target genes predicted in silico, we selected KLF6 as a candidate target gene for finding its role in milk fat metabolism in BMECs of Chinese Holstein cattle. The relative mRNA expression of KLF6 showed that compared with the bta-miR-148a-shNC group, bta-miR-148a significantly down-regulated (p < 0.001) the KLF6 expression in bta-miR-148a-mimic transfected BMECs with 4-fold change (Figure 3). Meanwhile, the relative expression of KLF6 was significantly up-regulated (p < 0.001) in BMECs transfected with bta-miR-148a-inhibitor as compared to that of the bta-miR-148a-shNC transfected BMECs with 4-fold change (Figure 3).
In the 3’UTR of KLF6 gene mRNA, the bta-miR-148a binding site predicted in silico (Figure 4A, B) was further validated by the renilla luciferase assay. For this purpose, the pmiR-RB-REPORTTM-KLF6-3’-UTR-WT and pmiR-RB-REPORTTM-KLF6-3’-UTR-MUT were, respectively, co-transfected with bta-miR-148a-mimic into BMECs. The results showed that the relative luciferase activity in the KLF6-WT+bta-miR-148a-mimic group was significantly decreased (p < 0.001) compared to the KLF6-MUT+bta-miR-148a-mimic group. On the other hand, there was a significant increase (p < 0.001) in luciferase activity in the KLF6-MUT+bta-miR-148a-mimic group compared to the KLF6-WT+bta-miR-148a-mimic group with 4-fold change (p < 0.001; Figure 4C). The combined analysis of bioinformatic data and luciferase activity shows that bta-miR-148a directly targets the 3’-UTR sequence of KLF6 mRNA with off-target activity identified in the mutant 3’-UTR sequence.
Western blotting was carried out to check the effects of transfection of bta-miR-148a-mimic, inhibitor, and shNC on protein expression of KLF6. The findings reflected that the pattern of KLF6 protein expression in BMECs transfected with the bta-miR-148a-inhibitor was higher relative to the mimic and shNC group at a 2-fold change (Figure 5). Meanwhile, the KLF6 protein expression was lower in the bta-miR-148a-mimic group than in the inhibitor and shNC group at 2-fold change. In addition to the mRNA expression of KLF6 and luciferase reporter assay, these findings further reveal that KLF6 is negatively regulated by bta-miR 148a.
The synthesis of triglycerides in BMECs of Chinese Holstein cow specifically influences milk fat composition. TG content analysis of transfected BMECs revealed a substantial (p < 0.05) decrease in the TG content of BMECs in the bta-miR-148a-mimic group relative to the bta-miR-148a shNC group. Contrarily, TG contents in BMECs transfected by the bta-miR-148-a-inhibitor were substantially higher (p < 0.01) than that of the bta-miR-148a-shNC transfect BMECs (Figure 6A). The analysis of cholesterol content in transfected BMECs showed that the cholesterol content in BMECs of the bta-miR-148a-mimic group was low compared to the bta-miR-148a shNC group. However, significantly high levels of cholesterol expression were found in bta-miR-148a-inhibitor transfected BMECs (p < 0.01) compared to bta-miR-148a-shNC transfected BMECs with a 4-fold change (Figure 6B). These findings suggest that bta-miR-148a negatively regulates TG and CHO contents in BMECs.
After the vectors of the pBI-CMV3-KLF6 gene were transfected into the cells, the total RNA of the cells was extracted for reverse transcription and Real-Time Quantitative Reverse Transcription PCR (qRT-PCR). As shown in Figure 7A, compared with the control group pBI-GFP-Neo-CMV3, the relative mRNA expression of KLF6 in pBI-CMV3-KLF6 was significantly higher (p < 0.001). For the KLF6 down-regulation study, three interference vectors of the KLF6 gene were transfected into the cells, respectively, and the total RNA of the cells was extracted for reverse transcription and qRT-PCR. The results showed that upon a comparison with the pb7SK-GFP-Neo, the expression of KLF6 in pb7SK-KLF6-siRNA1 was significantly lower (p < 0.01) with a 4-fold change, showing the best interference efficiency. Meanwhile, the relative mRNA expression in pb7SK-KLF6-siRNA2 and pb7SK-KLF6-siRNA3 was not significantly lower than the pb7SK-GFP-Neo. Therefore, the interference vector pb7SK-KLF6-siRNA1 was used for subsequent experiments (as shown in Figure 7B).
At 80% confluency, the BMECs were transfected with pBI-CMV3-KLF6 and pb7SK-KLF6-siRNA1 vectors, separately. After 24 h, the cell morphology and GFP expression were observed by fluorescence microscopy. The morphology of cells remained the same after transfection, and the transfection of pBI-CMV3-KLF6, pBI-GFP-Neo-CMV3, pb7SK-KLF6-siRNA1, and pb7SK-GFP-Neo was successful with over 60-70% transfection efficiency, which was ample to be used for subsequent experiments (Figure 8).
TG contents in BMECs transfected by pBI-CMV3-KLF6 were substantially higher (p < 0.001) than that of the pBI-GFP-Neo-CMV3 transfected BMECs (Figure 9A). These findings suggest that the overexpression of KLF6 positively regulates TG contents in BMECs. Meanwhile, TG contents in BMECs transfected by pb7SK-KLF6-siRNA1 were substantially lower (p < 0.01) than that of the pb7SK-GFP-Neo transfected BMECs (Figure 9B), suggesting that the down-regulation of KLF6 through pb7SK-KLF6-siRNA1 negatively regulates TG contents in BMECs. The TG contents in BMECs transfected by pBI-CMV3-KLF6 were substantially higher (p < 0.01) than those of the pb7SK-KLF6-siRNA1 as a 4-fold change (Figure 9C). These findings suggest that the overexpression of KLF6 increased the TG content compared to pb7SK-KLF6-siRNA1.
The analysis of cholesterol content in BMECs showed that the cholesterol content in pBI-CMV3-KLF6 transfected BMECs was significantly higher (p < 0.01) than that in the pBI-GFP-Neo-CMV3 transfected BMECs (Figure 10A). This evidence indicates that pBI-CMV3-KLF6 positively regulates the cholesterol contents in BMECs. The analysis of cholesterol content in transfected BMECs showed that the cholesterol content in BMECs transfected with pb7SK-KLF6-siRNA1 was significantly lower in comparison to the levels of cholesterol found in pb7SK-GFP-Neo BMECs (p < 0.001) compared to transfected BMECs (Figure 10B). This evidence indicates that pb7SK-KLF6-siRNA1 negatively regulates the cholesterol contents in BMECs. The cholesterol contents in BMECs transfected by pBI-CMV3-KLF6 were substantially higher (p < 0.01) than those of the pb7SK-KLF6-siRNA1 as a 4-fold change (Figure 10C). These findings suggest that the overexpression of KLF6 increased the CHO content compared to the pb7SK-KLF6-siRNA1.
Western blotting was used to assess KLF6 protein overexpression and down-regulation in BMECs. The findings showed that the pattern of KLF6 protein expression in BMECs transfected with the pBI-KLF6-CMV3 was higher relative to the pBI-GFP-Neo-CMV3 group (Figure 11A). The effects of transfection of pb7sk-KLF6-siRNA1 and pb7sk-GFP-Neo on the protein expression of KLF6 reflected that the pattern of KLF6 protein expression in BMECs transfected with the pb7sk-GFP-Neo was higher relative to the pB7sk-KLF6-siRNA1 group as a 2-fold change (Figure 11B). A similar trend of KLF6 expression in these groups was seen at the mRNA level, and these results further verified this trend of KLF6 expression at the protein level (Figure 11B).
The KEGG pathway (https://www.genome.jp/pathway/bta04152 (accessed on 1 January 2022) reported that AMPK negatively regulates the mTOR associated with the PPARG pathway (Figure 12A). The BMECs were cultured in 6-well plates. After 24 h, when the cells reached 90–95% growth, the total RNA was extracted and reverse transcribed, and qRT-PCR was performed. The relative mRNA expression of KLF6-KO-BMECs suggested that the expression of the AMPK was notably up-regulated (p < 0.001) in KLF6-KO-BMECs compared with the normal BMECs. Meanwhile, the mRNA expression of the mTOR was down-regulated in KLF6-KO-BMECs compared with the normal BMECs (Figure 12B), and the mRNA expression of the PPARG suggested that the expression of the PPARG in the KLF6-KO-BMECs was significantly down-regulated (p < 0.001) compared to the normal BMECs with the 4-fold change between the two groups. The mRNA expression results strongly suggested that KLF6 through AMPK/mTOR targets the PPARG pathway and plays a milestone in controlling lipid synthesis in BMECs. These results also have the same trend, which elucidates the KEGG pathway.
The BMECs were cultured in six-well plates. After 24 h, when the cells reached 90–95% growth, the protein was extracted, and a Western blot was performed. The relative protein expression of KLF6-KO-BMECs demonstrated that the AMPK was highly up-regulated (p < 0.01) in KLF6-KO-BMECs than in the normal BMECs. Meanwhile, the protein expression of the mTOR was down-regulated (p < 0.01) in KLF6-KO-BMECs compared with the normal BMECs (Figure 13A). In contrast, PPARG protein expression revealed that PPARG expression was down-regulated in KLF6-KO-BMECs relative to normal BMECs. The protein results strongly suggested that KLF6 through AMPK/mTOR targets the PPARG pathway and plays a milestone in controlling fat synthesis in BMECs. These results also have the same trend, which elucidates the KEGG pathway.
The KEGG pathway (https://www.genome.jp/pathway/bta03320 (accessed on 1 January 2022)) reported the PPARA pathway (Figure 14A). The BMECs were cultured in six-well plates. After 24 h, when the cells obtained 90–95% growth, the RNA was extracted and reverse transcribed, and qRT-PCR was performed. The relative mRNA expression of KLF6-KO-BMECs showed that the PPARA was significantly down-regulated (p < 0.001) in KLF6-KO-BMECs compared with the normal BMECs. So, the PPARA pathway genes, which are involved in lipogenesis and cholesterol metabolism, were also investigated, finding that the relative mRNA expression of the SCD was down-regulated (p < 0.001) in KLF6-KO-BMECs compared with the normal BMECs. Meanwhile, the mRNA expression of the MEI, CYPA1, CYP27A1, and LRX suggested that the expression of the MEI, CYPA1, CYP27A1, and LRX in the KLF6-KO-BMECs was significantly down-regulated compared to the normal BMECs with 4-fold change. The expression of the PPARA pathway genes was significantly down-regulated in the KLF6-KO-BMECs, which validates the importance of the KLF6 gene. The mRNA results strongly suggested that KLF6 targets the PPARA pathway, and PPARA has a milestone role in regulating fat synthesis in BMECs. These results elucidate that the KLF6 through the PPARA pathway regulates lipid synthesis in BMECs, as shown in Figure 14B.
The BMECs were cultured in six-well plates. After 24 h, when the cells attained 90–95% growth, the total protein was extracted, and the Western blot was performed. The protein expression of KLF6-KO-BMECs showed that the PPARA was significantly down-regulated (p < 0.001) in KLF6-KO-BMECs compared with the normal BMECs. Meanwhile, the protein expression of the SCD, MEI, CYPA1, CYP27A1, and LRX was significantly down-regulated in KLF6-KO-BMECs compared with the normal BMECs. The Western blot results strongly suggested that KLF6 targets the PPARA pathway and has a milestone role in controlling lipid production in BMECs. These results elucidate that KLF6 through the PPARA pathway regulates the lipid synthesis in BMECs, as shown in Figure 15A.
The BMECs were cultured in six-well plates. After 24 h, when the cell growth was about 90–95%, the total RNA was extracted and reverse transcribed, and qRT-PCR was performed. The relative mRNA expression of KLF6-KO-BMECs showed that the OXSM, FASN, MCAT, ABCG1, and GPAM were significantly up-regulated in KLF6-KO-BMECs as compared with the normal BMECs, which is shown in Figure 16A. Meanwhile, the mRNA expression of KLF6-KO-BMECs showed that the expression of the CBP4, HSD17B8, ACACA, ACACB, and AGPA was significantly down-regulated in KLF6-KO-BMECs as compared with the normal BMECs with 4-fold change, as shown in Figure 16B. These results strongly suggested that KLF6 plays an important role in lipid synthesis by targeting the marker genes related to lipid synthesis.
The string interaction shows that the different genes are interlinked with each other. Figure 17A string interaction showed that (https://stringdb.org/cgi/network?taskId=bNDo5kjnEDxr&sessionId=b8xaTD42Ytio (accessed on 10 February 2022)) the OXSM was regulated through the interaction of the FASN, MCAT, and CBP4 genes. Meanwhile, the main enriched KEGG pathways of OXSM showed that it regulates the fatty acid metabolism, biosynthesis, biotin metabolism, and biosynthesis of cofactors. In addition, the qPCR results showed that the KLF6 highly influences these genes and regulates these functions. Figure 17B of string interaction showed that (https://www.stringdb.org/cgi/network?taskId=b062fojmS7GI&sessionId=bw1BF3kKjtMh (accessed on 10 February 2022)) the SCD was regulated through the exchange of the ACACA, FASN, and HSD17B12 genes. Meanwhile, the main enriched KEGG pathways of SCD showed that it involves regulating the biogenesis of unsaturated fatty acid, metabolic pathway, fatty acid metabolism, and the PPAR and AMPK signaling pathway. The qPCR results showed that KLF6 highly influences these genes and regulates these functions. This string interaction and qPCR strongly suggested that KLF6 regulates the different lipogenesis pathways by targeting different genes.
Role of bta-miR-148a in milk fat synthesis in BMECs: With the advancement of biotechnology, numerous researchers have validated the pivotal role of miRNAs in cattle, which are involved with protein synthesis, mammary gland growth, and milk fat synthesis [45]. We aimed to investigate the role of bta-miR-148a and KLF6 in milk fat metabolism, primarily on the contents of TG, and CHO in BMECs. To this end, the BMECs were transfected with bta-miR-148a-mimic, inhibitor, and shNC. The presence of binding sites in 3’UTR of KLF6 mRNA that was found through bioinformatic analysis and verified through dual luciferase assay justified the negative correlation between bta-miR-148a and KLF6. To further explore the effects of bta-miR-148a on lipid metabolism in BMECs, the BMECs were transfected with bta-miR-148a mimic, inhibitor, and shNC. The results show that bta-miR-148a negatively regulates the expression of KLF6. The effect of bta-miR-148a transfection on TG and CHO contents in three experimental groups indicated that TG and CHO contents were low in the bta-miR-148a mimic group as compared to the inhibitor the and control group, while the TG and CHO content in the inhibitor group was high compared to the mimics and control group. Our result was similar to the study mentioned earlier conducted on mice: the deletion of the miR-148 harms hepatic cancer and lipid metabolism by targeting different genes, including ABCA1, and PGC1α [46]. When the mouse with KO miR-148a germline was used, it was revealed that miR-148a plays an important function in hepatic metabolism [17]. Another research identified that miR-148a has a significant role in immunity and epigenetic regulation [17,26]. The miR-148a, miR-186, and miR-200a are three of the most important curial miRNAs that play a crucial role in lactation and milk yield [17]. As a consequence of the association of miR-148 with the transcriptional factor MAFB, miR-148a has a key role in the differentiation of monocytes into osteoclasts [20]. In another investigation, it was shown that miR-148a, through inhibiting the DNMT1, plays an important role in the differentiation and maturation of liver cells [21]. Simultaneously, another study suggested that the SNP of miR-148 is linked with obesity. Previous research on marsupial tammar wallaby validated the most copious miRNAs during the lactation stage in the tammar wallaby, including miRNAs of the let-7 family (7f, 7a, and 7i), miR-148, miR-181, miR-184, miR-191, and miR-375 [47]. To this end, the results of this study coincide with the previous research and provide additional information which validates the role of bta-miR-148a in negatively regulating the TG and CHO content in milk fat synthesis in BMECs. Role of KLF6 in milk fat synthesis in BMECs: To delineate the role of KLF6, the KLF6 overexpression, and down-regulation was achieved through the transfection of BMECs with pBI-CMV3-KLF6 and pb7sk-KLF6-siRNA1, respectively. The result of KLF6 overexpression in BMECs through the transfection of pBI-CMV3-KLF6 showed a positive correlation with TG and CHO contents. This explains that through the overexpression of KLF6 in BMECs, the contents of TG and CHO were increased compared to the control group. For cross-checking, the KLF6 was down-regulated in BMECs through transfection with pb7sk-KLF6-siRNA1. The results showed that through the interference of KLF6 expression, the contents of TG, and CHO were decreased compared to the control group. Overall, these results highlight the importance of KLF6 in the synthesis of TG and CHO contents in BMECs, which can be generalized to the role of KLF6 in milk fat metabolism. KLF6 promotes the development of pre-adipocytes into adipocytes by modulating the expression of delta-like non-canonical notch ligand 1 (DLK1), which is a gene that suppresses adipocyte differentiation according to our results. The overexpression of KLF6 promoted adipocyte differentiation, while its silencing inhibited adipogenesis in 3T3-L1 cells [48]. Meanwhile, the role of the KLF6 gene in beef cattle study suggested that the KLF6 gene might be exploited as a viable marker gene for improving beef breeds through the marker-assisted selection of Qinchuan cattle [49]. In this study, the researcher validated that KLF6 plays a milestone role in enhancing meat quality in Qinchuan cattle [49]. According to another study, a single KLF6 allele reduced the risk of prostate cancer by 77% in human subjects highlighting the tumor suppressor attribution of KLF6 [50]. Meanwhile, another study reported that the amplification of the second exon of the KLF6 gene reveals the three SNP loci at 3332C > G; 3413C > T, and 3521G > A, which were found to be linked with more excellent body and carcass measurements in cattle [49]. Another study also reported the role of KLF6 in hepatic FA and lipid metabolism by inducing the expression of PPARG and TG accumulation in HepG2 cells. The KLF6 expression and TG accumulation were increased when HepG2 cells were treated with palmitic acid for imitating conditions such as a fatty liver, while the reverse happened upon the silencing of KLF6 [51]. The results of this study align with previous studies for the increased contents of TG in BMECs with the forced expression of KLF6 and decreased contents of TG in BMECs with the silencing of KLF6. KLF6 target the PPARA and PPARG: We determined whether KLF6 affects the AMPK/mTOR/PPARG pathway. The qPCR and WB results elucidated the expression of the AMPK up-regulated in the KLF6-KO-BMECs compared to the normal BMECs. Meanwhile, the expression of the mTOR/PPARG was decreased in the knock-out BMECs compared to the normal BMECs. From qPCR and Western blot, it confirmed that KLF6 through the AMPK/mTOR/PPARG pathway plays a significant role in the lipid synthesis in the BMECs. These results are also similar to the KEEG AMPK/mTOR/PPARG pathway. The previous study on PPARG reported that in early fat differentiation, PPARG is used as a master gene and plays a significant role in fat differentiation [52]. According to other studies in mice, PPARG isoform 2 is critical for hepatocyte lipid accumulation and controls the expression of lipogenesis and adipogenesis associated genes [53]. KLF6 activates the PPARG gene, which increases the amount of TG in the body [51,54]. The PPARG gene is a marker gene that plays an important role in early fat differentiation [52,55]. So, the previous research validates that PPARG is a key pathway that governs lipogenesis. The results of this study also corroborate with previous studies as the PPARG expression was significantly down-regulated in the KLF6-KO-BMECs compared to the normal BMECs. Overall, these results highlight the importance of the KLF6 gene in fat synthesis through PPARG pathway. In contrast to the normal BMECs, the KLF6 knock-out cell line showed a markedly reduced expression of PPARA and their selected genes in terms of mRNA and protein, suggesting that these genes are involved in lipogenesis and cholesterol metabolism. Preliminary studies have demonstrated an important role for PPAR in controlling the transcriptional activity of several key adipocyte-related genes, including aP2, C/EBP, GLUT4, and perilipin in response to insulin [56,57,58,59,60,61,62]. Meanwhile, another study in the mitochondria reported that PPARA regulates the fatty acid in the muscle [63,64], and the liver regulates fatty acid by targeting the a-carnitine palmitoyl transferase I genes [65]. The other study elucidated whether the PPARA-deficient mice fed with high fat accumulated the lipid in the liver, which showed that the PPARA has a major role in lipid metabolism [66]. So, the previous research validates that PPARA is a marker pathway that regulates lipogenesis. The findings of this study also coincide with previous studies as the expression of PPARA and its related genes was considerably down-regulated in KLF6-KO-BMECs as compared to the normal BMECs. During the conduct of this research, the role of the KLF6 gene in fat synthesis was demonstrated, which is controlled by the PPARA and PPARG pathways. The KLF6 regulates the PPARA and PPARG pathways, and the other marker genes significantly regulate fat synthesis. The qPCR of the marker genes elucidates the relative mRNA expression of KLF6-KO-BMECs and showed that the expression of OXSM, FASN, MCAT, ABCG1, and GPAM was significantly up-regulated in KLF6-KO-BMECs as compared with the normal BMECs. Meanwhile, the relative mRNA expression of KLF6-KO-BMECs showed that the expression of CBP4, HSD17B8, ACACA, ACACB, and AGPA was significantly down-regulated in KLF6-KO-BMECs as compared with the normal BMECs. The string interaction validates that these genes are interlinked with each other for the regulation of the function. The research results obtained via genome-wide-association, functional genomics, and comparative genomics analyses indicate that genes including SCD, DGAT1, ABCG2, and FASN were reported as genetic markers and candidate genes in milk fat traits [67]. Target interruption of the LXRα gene in mice reported lacking the expression of many fatty acid genes, for example, ACC, FAS, SREBP-1c, and SCD-1 [68]. Meanwhile, another study reported that the GPAM gene has an important role in enhancing the triglyceride content in the BFF cell line [69]. Further study validates that the GPAM gene has an important role in enhancing the triglyceride content in bovine mammary epithelial cells [70]. In contrast, the study on the ACACB gene reported that the SNP of the ACACB gene is associated with fatty acid regulation in milk [71]. Moreover, the study on ACACA elucidates that ACACA has increased the milk yield in dairy cattle [72]. Another study reported that the HSD17B8 gene has a key role in improving beef quality. The previous research validates the role of these genes in fat synthesis and lipid metabolism. Further, string interaction and the KEEG pathway validated that KLF6 is the major gene that regulates lipid synthesis in BMECs by targeting the PPARG, PPARA, and their other marker genes of fat synthesis. In this regard, this study provides additional information for validating the Importance of the PPARA and PPARG genes in lipid synthesis. The KEGG, previous research, and string interaction validate KLF6, by targeting PPARA and PPARG pathways, and the other fat marker genes, regulate fat synthesis in BMECs. Thus, in light of the previous research, in silico analysis, and experimental results of this study, we conclude that KLF6 is a potential candidate for increased milk and the synthesis of valuable milk components in dairy cattle through marker-assisted selection in molecular breeding. Interestingly, we have also revealed the role of bta-miR-148a and its target gene, KLF6, in the lipid metabolism of BMECs. However, further validation of these findings through in vivo experimentation would help the animal breeder's selective breeding. These results could be accounted for to improve the quality of dairy milk and the selection and breeding of dairy cows with the potential to produce better-quality milk.
We identified that bta-miR-148a negatively regulates the TG and CHO accumulation in BMECs. The KLF6 gene is a novel direct target of bta-miR-148a, which plays an important role in lipid metabolism in BMECs. The expression of KLF6 is directly linked to the contents of TG and CHO in BMECs. The bioinformatic and in vitro experimental results validated that KLF6, by targeting PPARA, the PPARG pathway, and the other fat marker genes, regulate fat synthesis in BMECs. Thus, in light of the results, we conclude that bta-miR-148a and KLF6 are potential candidates for increased milk and the synthesis of valuable milk components in dairy cattle through marker-assisted selection in molecular breeding. | true | true | true |
|
PMC9602169 | Shudai Lin,Li Qiu,Keying Liang,Haibo Zhang,Mingjian Xian,Zixi Chen,Jinfen Wei,Shuying Fu,Xiaocheng Gong,Ke Ding,Zihao Zhang,Bowen Hu,Xiquan Zhang,Yuyou Duan,Hongli Du | KAT2A/E2F1 Promotes Cell Proliferation and Migration via Upregulating the Expression of UBE2C in Pan-Cancer | 08-10-2022 | lysine acetyltransferase 2A (KAT2A),E2F transcription factor 1 (E2F1),ubiquitin conjugating enzyme E2 C (UBE2C),cell proliferation,cell migration,cell cycle,pan-cancer | Various studies have shown that lysine acetyltransferase 2A (KAT2A), E2F transcription factor 1 (E2F1), and ubiquitin conjugating enzyme E2 C (UBE2C) genes regulated the proliferation and migration of tumor cells through regulating the cell cycle. However, there is a lack of in-depth and systematic research on their mechanisms of action. This study analyzed The Cancer Genome Atlas (TCGA) to screen potential candidate genes and the regulation network of KAT2A and E2F1 complex in pan-cancer. Quantitative real-time PCR (qRT-PCR) and Western blotting (WB), cell phenotype detection, immunofluorescence co-localization, chromatin immunoprecipitation assay (ChIP), and RNA-Seq techniques were used to explore the functional of a candidate gene, UBE2C. We found that the expression of these three genes was significantly higher in more than 10 tumor types compared to normal tissue. Moreover, UBE2C was mainly expressed in tumor cells, which highlighted the impacts of UBE2C as a specific therapeutic strategy. Moreover, KAT2A and E2F1 could promote cell proliferation and the migration of cancer cells by enhancing the expression of UBE2C. Mechanically, KAT2A was found to cooperate with E2F1 and be recruited by E2F1 to the UBE2C promoter for elevating the expression of UBE2C by increasing the acetylation level of H3K9. | KAT2A/E2F1 Promotes Cell Proliferation and Migration via Upregulating the Expression of UBE2C in Pan-Cancer
Various studies have shown that lysine acetyltransferase 2A (KAT2A), E2F transcription factor 1 (E2F1), and ubiquitin conjugating enzyme E2 C (UBE2C) genes regulated the proliferation and migration of tumor cells through regulating the cell cycle. However, there is a lack of in-depth and systematic research on their mechanisms of action. This study analyzed The Cancer Genome Atlas (TCGA) to screen potential candidate genes and the regulation network of KAT2A and E2F1 complex in pan-cancer. Quantitative real-time PCR (qRT-PCR) and Western blotting (WB), cell phenotype detection, immunofluorescence co-localization, chromatin immunoprecipitation assay (ChIP), and RNA-Seq techniques were used to explore the functional of a candidate gene, UBE2C. We found that the expression of these three genes was significantly higher in more than 10 tumor types compared to normal tissue. Moreover, UBE2C was mainly expressed in tumor cells, which highlighted the impacts of UBE2C as a specific therapeutic strategy. Moreover, KAT2A and E2F1 could promote cell proliferation and the migration of cancer cells by enhancing the expression of UBE2C. Mechanically, KAT2A was found to cooperate with E2F1 and be recruited by E2F1 to the UBE2C promoter for elevating the expression of UBE2C by increasing the acetylation level of H3K9.
Many cancer treatment strategies have been developed based on targeting specific molecules related to specific gene mutation or specific gene expression, and breaking progress is being made with these cancer treatment strategies, and the potential is infinite [1,2,3]. Therefore, there is an urgent need to identify efficient therapeutical targets in tumors. Currently, many common molecular mechanisms across pan-cancer have been discovered by various studies [4,5,6]. High-throughput transcriptome was an important and effective data to identify the candidate targets or pathways [7,8,9]. Although tumors have great heterogeneity, there are commonalities among different types of tumors, so using the transcriptome data of large clinical pan-cancer samples to authenticate potential candidate targets upstream may be a good option. However, whether it will shed light on future cancer treatment still needs further comprehensive and in-depth research on pan-cancer. Previous studies have shown that there were abnormalities of histone acetylation modification in various cancers, including liver, lung, and breast [10,11]. Histone acetylation mediates the expression and activation of genes related to cell proliferation, differentiation, and apoptosis, which could affect the occurrence and development of tumors [12,13] Histone lysine acetyltransferases (KATs) and histone deacetylases (HDACs) are key effectors balancing between histone acetylation and deacetylation. KAT2A, the first discovered KAT that was related to transcription, has been reported to be involved in gene transcription, DNA repair, nucleosome assembly, and cell cycle regulation in pan-cancer. Additionally, it was significantly up-regulated in many cancers to promote the growth of tumor cells/cell proliferation, and the invasion and migration of cancer cells [14,15,16,17,18,19,20]. Therefore, KAT2A may be a significant oncological target with effective therapeutics for several cancers. E2F1 was found as a key regulator of G1/S transition, and to promote the transcription of plenty of critical genes for cell-cycle progression [21]. Many reports have found that E2F1 played a central role in cancer development, such as in breast cancer [22,23], bladder cancer [24], and prostate cancer [25]. It is suggested that the up-regulation of E2F1 can promote the proliferation, migration, and invasion of these cancer cells, and it is also significantly related to the clinical stage of different cancer types, the depth of tumor invasion, as well as the metastasis and lesion size of lymph nodes [26]. As a crucial catalytic component of transcription regulation complex, it has been suggested that KAT2A can increase the chromatin accessibility of transcription factors (such as E2F1) and form protein complexes with them. Moreover, it could be recruited to the promoter regions of genes involved in the cell cycle, DNA damage repair, and cell migration, consequently enhancing their expression through increasing the acetylation level of H3K9 on these gene-promoting regions [27,28]. For example, KAT2A has been explored to cooperate with E2F1 and be recruited by E2F1 to the promoters of cyclin D1 and cyclin E1 [16], and it is amplified in breast cancer 1 genes [29]. UBE2C is a member of the E2 ubiquitin-conjugating enzyme family [30]. Ubiquitination is an important cellular mechanism for targeting proteins for degradation, which is involved in numerous cell processes, such as cell cycle progression, antigen presentation, transcription, and programmed cell death [31]. Many studies have shown that the expression of UBE2C is upregulated in a variety of human malignancies, such as tongue squamous cell carcinoma [32], breast cancer [33], endometrial cancer [34], melanoma [35], and rectal carcinoma [36]. All these findings suggested that UBE2C is closely associated with the development of cancer, and could be used as a potential therapeutic target for different types of cancers. Taken together, these reports indicated that KAT2A, E2F1, and UBE2C play a fundamental role in the progression of several types of cancers. Hitherto, no conclusive study has reported the role of KAT2A and E2F1 interactions in the pan-cancer landscape. With such conspicuous roles of both KAT2A and E2F1 in cellular functions and putative links to cancer, we investigated the common molecular mechanisms and potential transcription target of their interaction across pan-cancer. This study revealed the common characteristics of KAT2A/E2F1/UBE2C and clarified the mechanism of this axis across pan-cancer through RNA-seq dada and in vitro experiments, which might shed light on pan-cancer treatment.
The expression data and corresponding clinical information of different kinds of cancer patients were downloaded from The Cancer Genome Atlas (TCGA). The GSE137172 set was downloaded from the Gene Expression Omnibus database (GEO; http://www.ncbi.nlm.nih.gov/geo/, accessed on 13 January 2021) for analyzing differentially expressed genes (DEGs) after knocking down UBE2C.
KAT2A, E2F1, and UBE2C were formed with the Transcripts Per Million (TPM) mean of each mRNA expression of samples in TCGA, and the exact sample sizes of cancer and normal samples used are reported in Table S1. The expression profiles of KAT2A, E2F1, and UBE2C were analyzed using GEPIA (Gene Expression Profiling Interactive Analysis, http://gepia.cancer-pku.cn/, accessed on 14 April 2021) online analysis [37]. Then, comparisons between tumor and normal tissues were analyzed. The relative expression levels of KAT2A, E2F1, and UBE2C to ACTB were also analyzed using corresponding cancer cell lines in the Cancer Cell Line Encyclopedia (CCLE) database, respectively.
The TPM expression of KAT2A, E2F1, and UBE2C at different pathological stages in the TCGA database were represented by box plots, and Student’s t-test was employed to compare the relative expression levels among different pathological stages. p < 0.05 indicated statistically significant differences.
The clinical outcome of patients with different types of cancers was determined using Kaplan–Meier survival curves. For the overall survival (OS), the samples were divided into two groups according to the median expression of the mRNAs (high vs. low). With the use of R packages (survival, version 3.2.7; survminer, version 0.4.8), Kaplan–Meier survival analysis and the log-rank test were employed to compare OS between the tumor and normal cohorts. p < 0.05 indicated statistically significant differences.
The samples of different types of cancers in the TCGA databases were separated into 30% each of KAT2A, E2F1, and UBE2C high and low groups to obtain DEGs using the “DESeq2” package (version 1.28.1) in R language (version 4.0.2). |Fold Change| > 1.5 and FDR < 0.05 were set as the statistical threshold value of DEGs. Using the transcriptome data of 11 tumor types (with normal tissues more than 30) in TCGA, which contains the significantly highly expressed level of KAT2A, overlapping DEGs were screened according to the following conditions: each tumor KAT2A and E2F1 were grouped according to the 30% high and low groups to obtain: the DEGs (FC > 1.2, FDR < 0.05, KAT2A 30%-Up, and E2F1 30%-Up), respectively; the correlation coefficient with KAT2A and E2F1 > 0.3, respectively; and DEGs that were highly expressed in tumor tissues compared with normal tissues.
Spearman’s correlation coefficient analysis was performed to explore the correlation among KAT2A, E2F1, and UBE2C in 11 out of 33 cancers with more than 200 tumor tissues from the TCGA database, and the corresponding cell lines in the CCLE database.
Functional enrichment analysis, gene ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were conducted by the R package (clusterProfiler, version 3.16.1) to explore the different molecular mechanisms and involved pathways between high and low UBE2C expression. The protein–protein interaction (PPI) network of DEGs was obtained from the STRING (version 11.0) database [38].
The results from the TCGA RNA transcriptome data and the existing GEO dataset were combined to obtain the overlapping DEGs, and the key genes or target proteins and signal pathways regulated by UBE2C were searched. The common mechanism of KAT2A/E2F1/UBE2C affecting tumor cell proliferation, cell cycle, and apoptosis in different tumors were explored through GO and KEGG function enrichment analyses.
The single-cell RNA-Seq data were analyzed as described previously [39] Expression data were extracted from a previous study [39] and violin plots were drawn using R.
Parental MCF-7 breast cancer, NCI-H460 large cell lung carcinoma HepG2 liver cancer, and BxPC3 pancreatic cancer cell lines were gifts from Dr. Peng Huang, Sun Yat-sen University Cancer Hospital, Guangzhou, China. The 786-O renal clear cell carcinoma cell line was purchased from Cell Resource Center, Shanghai Academy of Biological Sciences, Chinese Academy of Sciences. MCF-7, HepG2, and BxPC3 were cultured in DMEM medium with 10% fetal bovine serum, penicillin (100 U/mL), and streptomycin (100 U/mL) at 37 °C in air with 5% CO2. MCF-7 was cultured with 0.2 mg/mL insulin. NCI-H460 and 786-O were in RPIM-1640 medium with 10% fetal bovine serum, penicillin (100 U/mL), and streptomycin (100 U/mL) at 37 °C in air with 5% CO2. For transient knockdown studies, KAT2A-shRNA, UBE2C-shRNA (Fitgene, Guangzhou, China), and control shRNA (shNC) plasmids, and a final concentration of 60 nM of both E2F1-siRNA and control siRNA (siNC) (Hanheng, Shanghai, China) (Table S2) were transfected for 24 h according to Lipofectamine™ 3000 (Thermo Fisher Scientific, Waltham, MA, USA). NCI-H460 cells with a stable knockdown of KAT2A were established by transfection with a KAT2A-shRNA (shKAT2A-1) lentiviral vector.
Real-time quantitative PCR (qPCR) analysis was performed according to the user’s manual using the StepOnePlus™ Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) and Power SYBR Green PCR Master Mix (Applied Biosystems) kits. All samples were analyzed in triplicate, and the expression of KAT2A, E2F1, and UBE2C was normalized relative to that of GAPDH, which was used as an internal loading control. The primers for qPCR are listed in Table S2.
The whole-cell lysate or the immunocomplexes were separated by 8 to 12% SDS-PAGE and transferred onto a polyvinylidene difluoride (PVDF) membrane (Millipore, Billerica, MA, USA). Anti-KAT2A (Cell Signaling, Danvers, MA, USA, 1:1000, #3305), anti-E2F1 (Invitrogen, Waltham, MA, USA; Thermo Fisher Scientific, Waltham, MA, USA, 1:1000, MA1-23202), anti-H3K9ac (Cell Signaling Technology, 1:1000, 9649S), anti-UBE2C (Invitrogen, Thermo Fisher Scientific, 1:1000, PA5-27223), anti-β-actin (Beyotime Biotechnology, Shanghai, China, 1:1000, AA128), and horseradish peroxidase (HRP)-conjugated secondary antibodies (anti-mouse and anti-rabbit IgG) (Beyotime Biotechnology, 1:2000, A0208, A0216) antibodies were used to detect each protein. Bands were detected using BeyoECL Star chemiluminescence substrate (Beyotime Biotechnology, P0018AM).
The detailed immunohistochemistry procedures were performed as described before [40]. Cells were seeded into the 35 mm laser confocal petri dishes under normal culture conditions to reach 60% density without any treatment to prepare for performing cell immunofluorescence. After incubating with KAT2A and E2F1 primary antibodies (Invitrogen, Thermo Fisher Scientific, 1:1000, MA5-14884, MA1-23202), the cells were then incubated with the corresponding diluted IgG fluorescent secondary antibodies (Invitrogen, Thermo Fisher Scientific, 1:5000, A11029, A11012) for 1 h at room temperature in the dark. Then, the cells were stained with nucleus DAPI dropwise and incubated for 5 min in the dark. The mounting solution containing anti-fluorescence quencher (Invitrogen, Thermo Fisher Scientific, P36971) was dropped, followed by observing and collecting the image under a fluorescence microscope (Wetzlar, Germany, Leica TCS SP8 X).
ChIP was performed according to the instructions of the Pierce Agarose ChIP Kit (Thermo Fisher Scientific, 26156). The E2F1 binding sites on the UBE2C promoter were analyzed using the JASPAR online tool (http://jaspar.genereg.net/, accessed on 8 June 2020). ChIP-qPCR data were shown as the percentage of input following normalization with no antibody (mock). The primers for ChIP-qPCR are listed in Table S2. Co-IP was performed using indicated antibodies and IgG (Invitrogen) according to the manufacturer’s instructions. In brief, cell lysates were incubated with antibody-conjugated beads at 4 °C for 2 h. Then, the beads were washed extensively and boiled in SDS loading buffer. A total of 4% of total protein was used per IP, about 50 µg/100 µg.
The luciferase assay was performed as described previously [41].
Cell proliferation assays were performed using the Cell Counting Kit-8 (CCK-8; Sangon Biotech, Shanghai, China) according to the manufacturer’s instructions. Briefly, cells were seeded onto 96-well plates (3 × 103 cells per well) and transfected when they reached 70–80% of confluence according to the protocol of Lipofectamine™ 3000, and were then added with 10 µL of CCK-8 solution and cultured for 1 h at 37 °C in air with 5% CO2 on designated days. The absorbance was measured at 450 nm using TECAN infinite M200 (Softmax Pro., Molecular Devices, Sunnyvale, CA, USA). For EdU assay, the cells were treated for 48 h, followed by using the BeyoClick™ EdU Cell Proliferation Kit with Alexa Fluor 594 (Beyotime Biotechnology, C0078S) according to the manufacturer’s protocol.
The stable KAT2A overexpression NCI-H460 cells were seeded in a 6-well plate with 2000 cells per well. After culturing for 14 days, the culture medium was discarded, was washed carefully with 1 × phosphate-buffered saline (PBS) (Gibco, Grand Island, NY, USA) once, and 1 mL of methanol solution was added to each well to fix the cells for 30 min. The methanol solution was aspirated, 1 mL of crystal violet stain solution was added to each well, and it was left at room temperature for 30 min. The crystal violet was recycled and each well was washed with distilled water, and the culture plate was placed upside down on absorbent paper to absorb the water. It was dried naturally, and pictures were taken using a digital camera. The number of clones with more than 10 cells under the microscope (4 × magnification) was counted. Finally, we calculated the clone formation rate = (number of clones/number of inoculated cells) × 100%.
A wound healing assay was performed to detect the migration of three kinds of cancer cell lines after treatment. Cells growing the in log phase were trypsinized and seeded in 24-well plates until confluent. A total of 1 × 105 cells per well were seeded in 24-well plates. After 24 h, the cells were transfected with shKAT2A-1, siE2F1-2, and shUBE2C-3, and the corresponding control shNC and siNC using LipofectamineTM 3000 (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. After transfection, cells were incubated at 37 °C and 95% confluent cells were used for a wound healing assay. Wounds were made using a 10 µL sterile tip. After incubation for 0, 24, 48, 72, and 96 h, the cells were photographed under an inverted microscope, respectively. The distance between the two edges of the scratch (wound width) was measured at 8 sites using ImageJ in each image (40× magnification). MCF-7, 786-O, and NCI-H460 cells were transiently transfected with shUBE2C-3 for 48 h, followed by using Falcon® Permeable Support for a 24-well Plate with 8.0 µm Transparent PET Membrane (Corning, NY, USA) for measuring cell migration and invasion. After taking pictures of the cells, we added 1 mL absolute ethanol. With sufficient shaking, the light absorbance was measured at 570 nm. The optical absorbance (OD) value was used to plot the dilution of the samples, and the curves of the standard product and the samples were compared.
The cells were treated with shUBE2C-3 and shNC for 48 h, respectively. Afterwards, the cells were digested with trypsin without EDTA. After termination, the cells were centrifuged at 1000 rpm for 5 min at 4 °C to remove the supernatant containing trypsin, and were then washed 2 times with precooled PBS. The cells were regenerated with serum-free medium, and cell apoptosis was analyzed using an Annexin V-FITC Apoptosis Detection Kit (Beyotime Biotechnology, C1062M) according to the instructions.
After 48 h treatment of shNC and shUBE2C, the cells were washed twice with precooled PBS, centrifuged at 1200 rpm 4 °C for 5 min, fixed with 70% ethanol, and then stained with 10 µg/mL propidium iodide in a solution containing 100 µg/mL RNase in PBS; the cell pellet was slowly and fully resuspended, and was incubated at 37 °C for 30 min in the dark. Then, the cells were analyzed on a BD FACSCalibur (Becton, Dickinson and Company, San Jose, CA, USA).
Total RNA was extracted from the cells transfected for 48 h using Trizol reagent (Takara, Japan) and stored at −80 °C. The complementary DNA (cDNA) libraries of each pooled RNA sample for single-end sequencing were generated using the NEBNext® UltraTM RNA Library Prep Kit for Illumina® (NEB, E7530L) according to the manufacturer’s instructions. The cDNA libraries were subjected to the NovaSeq 6000 system (Illumina), according to commercially available protocols. The changed RNAs were validated by quantitative PCR using the primers listed in Table S2.
The quality control of raw sequencing data was conducted using fastp [42] The clean reads from RNA-seq were aligned to the human reference genome sequence, GRCH38.p13, using the HISAT2 program (v2.2.3) [43] The gene expression level was determined by the featurecounts function of subread software with the genome annotation file from GENCODE (v36) [44,45,46].
Data were showed as the mean values with the standard error of the mean. Statistical differences were determined using Student’s t-test or one-way ANOVA. p < 0.05 was considered to indicate statistical significance.
From the GEPIA online analysis, we found that compared with the expression level distributed in a normal person, KAT2A was slightly higher in cancer patients, and E2F1 was obviously higher expressed in cancer patients (Figure 1A,C) [47,48]. As a result of the TCGA pan-cancer transcriptome data, the expression levels of KAT2A and E2F1 were significantly up-regulated in 16 and 20 kinds of tumor tissues compared with normal tissues, respectively. Both the expression levels of KAT2A and E2F1 were significantly up-regulated in the same 16 cancers, and E2F1 was also higher expressed in CESC, GBM, KICH, and UCEC cancers (Figure 1B,D). In addition, the expression levels of KAT2A and E2F1 in pathological stages were remarkably higher expressed in stage III and IV in 10 cancers (Figure 1E,F). These findings suggested that KAT2A and E2F1 may play an important role in tumor development. From the results of CCK-8 and wound-healing assays, the proliferation of MCF-7, 786-O, and NCI-H460 cells (Figure 2A–F), and the cell migration ability of NCI-H460 cells (Figure 2G,H) were suppressed with the knockdown of either KAT2A or E2F1, compared with the control group. Moreover, the knockdown of KAT2A significantly inhibited the colony-forming ability of NCI-H460 cells (Figure 2I).
In order to explore the potential target genes co-regulated by KAT2A and E2F1, the differentially expressed genes (DEGs) of KAT2A and E2F1 were screened according to the conditions. A total of 9 to 1202 overlapping DEGs were obtained from different cancer types (Figure S1A–K). Among 11 cancers, there were 222 genes that appeared in more than five cancer types at the same time (Figure S1L), including UBE2C, which appeared in six types of cancer (Table S3). The GO analysis of the overlapping 222 DEGs revealed that the biological pathways (BP) were related to DNA replication, nuclear division, the regulation of cell cycle phase transition and cell cycle checkpoint, and so on (Figure S2A); the enriched cell components (CC) included chromosomal region, chromosome, centromeric region, spindle, and kinetochore (Figure S2B); and the molecular function (MF) was involved in the action of catalytic activity, acting on DNA, deoxyribonuclease activity, and so on (Figure S2C). The KEGG pathway analysis indicated that these genes were significantly enriched in the cell cycle, DNA replication, homologous recombination, base excision repair and nucleotide excision repair, and so on (Figure S2D). From the result of correlation analysis in tumor tissue samples across 11 cancer types and cell line samples in CCLE, we found that the correlation coefficient between the two in KAT2A, E2F1, and UBE2C was 0.037–0.82 (Figure 3A). Considering the analysis results of the transcriptional data of various cancers in the TCGA database (Figure S1), it is suggested that UBE2C may be a downstream gene that is co-regulated by KAT2A and E2F1. Correspondingly, we found that both the mRNA and the protein levels of UBE2C were significantly suppressed after the knockdown of KAT2A or E2F1 in MCF-7, 786-O, and NCI-H460, respectively (Figure 3B–H). Moreover, the mRNA/protein changes of E2F1 were significantly inhibited after knocking down KAT2A (Figure 3B–D). Furthermore, the knockdown of UBE2C caused an extreme decrease of the KAT2A protein level (Figure 3H). Additionally, the results of cellular immunofluorescence showed the nuclear localization of KAT2A and E2F1 in MCF-7, 786-O, and NCI-H460 cells (Figure 4A). Moreover, ChIP-qPCR showed that KAT2A and E2F1 can bind to the −322 to +39 region of the UBE2C promoter (with more than one-fold enrichment), not −1081 to −819, or +11 to +215 regions (with less than one-fold enrichment) (Figure 4B–G). Additionally, the dual luciferase reporter gene experiment also confirmed that E2F1 could bind to the −273 to −266 region of the UBE2C promoter (Figure S3A–F). Moreover, the ChIP-qPCR assay demonstrated that the KAT2A bound to the promoter region of UBE2C, and increased the H3K9 acetylation level in this promoter region, which suggested that KAT2A and E2F1 may cooperate to regulate UBE2C gene transactivation via histone modification. Moreover, the Co-IP assays showed that KAT2A, E2F1, and H3K9ac could bind to each other (Figure S3G). These results demonstrated that KAT2A may promote the expression of UBE2C through combining with E2F1 to the E2F1 binding site on UBE2C promoter −322/+39 region to increase the acetylation level of H3K9, and consequently stimulated cancer cell proliferation and migration.
The results of the GEPIA online analysis and TCGA database showed that the expression of UBE2C was higher in many types of tumors than normal tissues, especially in brain, lung, and breast (Figure 5A,B). Additionally, the result of the expression levels of UBE2C in pathological stages showed that it was significantly highly expressed in stage III and IV in 10 types of cancer (Figure 5C). In addition, based on our previous study [39], the analysis of tumor tissue single-cell transcriptome data showed that the expression of UBE2C has a significant up-regulation trend in cancer cell clusters (CS), such as CS4 of CRC, CS2 of LC, CS4 of OV, CS3 of PDAC, and CS5 of SCC (Figure S4). Moreover, the survival analyses revealed that the high expression of UBE2C was significantly associated with a poor prognosis in patients of nine cancer types (Figure 5D). The above results suggested that UBE2C was up-regulated in a variety of cancer tissues/cells, and it may play a pivotal role in the development of pan-cancer.
To investigate the function of UBE2C in MCF-7, 786-O, and NCI-H460 cells, we transfected the cell lines with UBE2C shRNA (shUBE2C), and the qRT-PCR results showed that the expression of UBE2C in shUBE2C-treated group was significantly lower than that in the negative control group (shNC) (Figure 6A). Additionally, the results from the CCK-8 assay showed that the proliferation of different cancer cell lines was markedly suppressed with knocking down UBE2C in MCF-7, 786-O, and NCI-H460 cells compared with the control group (Figure 6B–D). As a key downstream gene of KAT2A and E2F1, we also found that UBE2C significantly inhibited the proliferation of the liver cancer cell line, HepG2, and pancreatic cancer cell line, BxPC3, through a CCK-8 assay (Figure S5), which suggested that UBE2C plays a critical role in pan-cancer. The EdU experiment suggested that the number of proliferating cells was significantly reduced after the knockdown of UBE2C at 48 h (Figure 6E). In addition, we found that the cell proliferation ability was significantly inhibited after co-interfering with KAT2A and UBE2C, E2F1, and UBE2C, compared to that of only knocking down KAT2A or E2F1 in NCI-H460 cells (Figure 6F,G). Moreover, compared with the control group (pLVX-puro), the overexpression of UBE2C (pLVX-UBE2C) significantly promoted cell proliferation, whereas knocking down KAT2A and E2F1 caused an inhibition on cell proliferation, which can be restored by the overexpression of UBE2C (Figure 6H,I). Moreover, the overexpression of KAT2A (pLVX-KAT2A) also significantly promoted cell proliferation, whereas knocking down UBE2C caused an inhibition on cell proliferation, which can be restored by the overexpression of KAT2A (Figure 6J). These results not only suggested that the downregulation of UBE2C could significantly inhibit the proliferation of tumor cells, but also confirmed that KAT2A and E2F1 could indeed affect the proliferation of tumor cells through regulating the expression of UBE2C. Next, we examined the effects of UBE2C on cell metastasis and apoptosis. The results of the wound-healing assay (Figure S6A–C) and transwell experiment (Figure S7A–C) showed that the downregulation of UBE2C remarkably suppressed the cell migration ability in MCF-7, 786-O, and NCI-H460 cells. Moreover, the number of apoptotic cells in MCF-7, 786-O, and NCI-H460 cells (the total number of cells in Q2 and Q3 in the figure) was significantly higher after 48 h of knockdown of UBE2C than that of the control groups (Figure S8A–C).
To better understand the molecular signatures after the knockdown of UBE2C in NCI-H460 and MCF-7 cells, we performed RNA-seq analysis and analyzed the DEGs. As a result, a total of 5539 and 1756 DEGs in NCI-H460 and MCF-7 cancer cells was screened with |FC| > 1.2 and FDR < 0.05, respectively (Figure 7A,B). As expected, the significantly enriched pathways of GO_BP of NCI-H460 were translational initiation, regulation of cell growth, cell cycle arrest, cell cycle checkpoint, and so on (p.adjust < 0.05) (Figure 7C). The KEGG pathways of NCI-H460 were the included ribosome, cell cycle, adherens junction, protein processing in endoplasmic reticulum, and apoptosis (Figure 7D). Additionally, the GO analysis showed that the DEGs of MCF-7 were highly associated with cell cycle G1/S phase transition, the intrinsic apoptotic signaling pathway, histone modification, and DNA replication initiation (Figure 7E). The KEGG analysis showed that the DEGs of MCF-7 were significantly related to protein processing in the endoplasmic reticulum, DNA replication, and the cell cycle (Figure 7F). These results suggested that UBE2C may promote the development of pan-cancer through influencing the cell cycle. Notably, there were 772 overlap DEGs in the two cell lines, including Cyclin Dependent Kinase Inhibitor 1B (CDKN1B), Inhibin Subunit β A (INHBA), Ras Homolog Family Member U (RHOU), BTG Anti-Proliferation Factor 1 (BTG1), Transducer Of ERBB2, 1 (TOB1), N-Myc Downstream Regulated 1 (NDRG1), MAX Network Transcriptional Repressor (MNT), MAX Interactor 1, Dimerization Protein (MXI1), and Ajuba LIM Protein (AJUBA), which have been reported to be involved in the cell cycle and proliferation [49,50,51,52,53,54,55,56]. Intriguingly, the GO analysis showed that the overlap genes were highly associated with the G1/S transition of the mitotic cell cycle, intrinsic apoptotic signaling pathway, and negative regulation of growth and DNA replication initiation (Figure 7G), and the KEGG analysis showed that they were significantly related to cell cycle and protein processing in the endoplasmic reticulum (Figure 7H). Moreover, the heat map showed that the DEGs in two cancer cell lines with interfering UBE2C was involved in the cell cycle (Figure 7I), suggesting that UBE2C participating in the cell cycle may be a common regulatory mechanism of pan-cancer development. Therefore, we performed cell cycle experiments and found that the cell cycle was significantly blocked in the G1 phase after the knockdown of UBE2C, suggesting that UBE2C may affect the cell cycle by regulating the G1/S phase transition (Figure S9). To further validate how UBE2C affects the progress of cancer development through the cell cycle, the DEGs were obtained based on the 30% of high and low expression levels of UBE2C in 11 types of cancer tissues. A total of 1514 DEGs, which appeared in five types of cancer, were screened. A total of 596 DEGs was obtained from 1514 DEGs overlapped with the DEGs from the GSE173127 dataset. The GO analysis of these 596 DEGs revealed that they were mainly involved in organelle fission, DNA replication, and the cell cycle checkpoint (Figure S10A–C), and the KEGG analysis showed that they were also involved in the cell cycle, DNA replication, and p53 signaling pathway (Figure S10D). The PPI analysis results of the above 596 DEGs showed that UBE2C can form an interaction network with Tumor-Transforming Protein 1 (PTTG1), Polo Like Kinase 1 (PLK1), Cyclin Dependent Kinase 1 (CDK1), Cell Division Cycle 20 (CDC20), cyclin B2 (CCNB2), cyclin B1 (CCNB1), cyclin A2 (CCNA2), and Breast and Ovarian Cancer Susceptibility Protein 1 (BRCA1) (Figure S10E). Genes regulated by UBE2C can also form an interaction network with TP53 (Figure S10F). Furthermore, hub genes contained kinesin family members 11 and 20A (KIF11, KIF20A), BUB1 mitotic checkpoint serine/threonine kinase (BUB1), Aurora Kinase B (AURKB), CCNA2, CCNB2, CDC20, CDK1, topoisomerase (DNA) II α (TOP2A), and DLG associated protein 5 (DLGAP5) (Figure S10G).
Multiple studies have revealed that KAT2A was highly expressed in a variety of cancers compared with adjacent tissues, such as liver cancer [57], colon adenocarcinoma tissues [58], and non-small cell lung cancer tissues [16]. The downregulation of KAT2A can significantly reduce the proliferation and migration of cancer cells and the growth of xenograft tumors [57,59]. In this study, we found that KAT2A was generally highly expressed in seven cancer tissues compared with normal tissues, including BLCA, CHOL, ESCA, and HNSC, KIRP, STAD, and THCA. As a transcription factor, high E2F1 levels were commonly associated with aggressive cancer and poor patient prognosis for multiple cancer types. E2F1 transcription factor was a key regulator of genes required for cell cycle progression, cell proliferation, and differentiation [21,60], and played a key regulatory role in the invasion–metastasis cascade of certain cancer types [61,62]. Previous studies have revealed that E2F1 can induce cell metastasis by inducing chemoresistance, angiogenesis, secondary site extravasation, and EMT [63,64,65,66,67,68,69] This study found that E2F1 was significantly highly expressed in 20 cancers, and it was significantly related to the poor prognosis in patients of 9 different cancers. By screening the transcriptome data of 11 cancer tissues of TCGA, it was exposed that the gene set potentially co-regulated by KAT2A and E2F1 was significantly enriched in the cell cycle, and participated in DNA replication, base excision repair, and nucleotide excision repair. It was suggested that the regulatory network of KAT2A and E2F1 involved in the cancer process was very complex, and it could regulate the expression level of genes involved in the cancer development process. Previous studies have shown that KAT2A could increase the chromatin accessibility of E2F1, DNA Damage Inducible Transcript 3 (DDIT3), and other transcription factors, and form protein complexes with them, and then be recruited to the promoter regions of related genes, consequently enhancing their expression through increasing the acetylation level of H3K9 on these gene-promoting regions and regulating the development of cancer. In this study, the ChIP-qPCR and Co-IP results have demonstrated that KAT2A may interact with E2F1 and form a complex to bind the UBE2C promoter in MCF-7, 786-O, and NCI-H460 cells. The complex could increase the level of H3K9ac, thereby promoting the expression of UBE2C. Moreover, we uncovered that the E2F1 binding site region (−322/+39) of the UBE2C promoter was consistent with the results of a previous study [70]. This study explored the regulatory relationship between KAT2A/E2F1 and UBE2C, and found that KAT2A can regulate the expression of UBE2C through interacting with E2F1. KAT2A can also affect the expression of E2F1, which is in agreement with previous studies [16,71]. E2F1 can also regulate the expression of KAT2A, but whether it affects the expression of KAT2A through the E2F1 binding site on the KAT2A promoter or other mechanisms still needs further in-depth research. Interestingly, the downregulation of UBE2C can inhibit KAT2A protein expression levels, indicating that UBE2C can regulate the expression of KAT2A and form feedback regulation. Previous studies have reported that the SCF-Cyclin F ubiquitin ligase complex participates in the ubiquitination and degradation of E2F1 [72], but whether its degradation required the participation of UBE2C has not been reported. If it was required for the involvement of UBE2C, it can be explained that the E2F1 protein was significantly upregulated in 786-O and NCI-H460 after the knockdown of UBE2C for the reason that the degradation of the E2F1 protein was reduced after the knockdown of UBE2C. However, what role these up-regulated E2F1 proteins may play need further research. Given the clinical and functional significance of KAT2A/E2F1/UBE2C in pan-cancer, we concluded that KAT2A/E2F1/UBE2C and its associated pathway were crucial for cancer carcinogenesis, and targeting this pathway may be pivotal in the prevention or treatment of pan-cancer. | true | true | false |
|
PMC9602374 | Francesco Passaretti,Laura Pignata,Giuseppina Vitiello,Viola Alesi,Gemma D’Elia,Francesco Cecere,Fabio Acquaviva,Daniele De Brasi,Antonio Novelli,Andrea Riccio,Achille Iolascon,Flavia Cerrato | Different Mechanisms Cause Hypomethylation of Both H19 and KCNQ1OT1 Imprinted Differentially Methylated Regions in Two Cases of Silver–Russell Syndrome Spectrum | 16-10-2022 | Silver–Russell syndrome,imprinting disorders,microduplication,differentially methylated regions,DNA methylation defects,11p15.5 imprinted genes cluster | Silver–Russell syndrome is an imprinting disorder characterised by pre- and post-natal growth retardation and several heterogeneous molecular defects affecting different human genomic loci. In the majority of cases, the molecular defect is the loss of methylation (LOM) of the H19/IGF2 differentially methylated region (DMR, also known as IC1) at the telomeric domain of the 11p15.5 imprinted genes cluster, which causes the altered expression of the growth controlling genes, IGF2 and H19. Very rarely, the LOM also affects the KCNQ1OT1 DMR (also known as IC2) at the centromeric domain, resulting in an SRS phenotype by an unknown mechanism. In this study, we report on two cases with SRS features and a LOM of either IC1 and IC2. In one case, this rare and complex epimutation was secondary to a de novo mosaic in cis maternal duplication, involving the entire telomeric 11p15.5 domain and part of the centromeric domain but lacking CDKN1C. In the second case, neither the no 11p15.5 copy number variant nor the maternal-effect subcortical maternal complex (SCMC) variant were found to be associated with the epimutation, suggesting that it arose as a primary event. Our findings further add to the complexity of the molecular genetics of SRS and indicate how the LOM in both 11p15.5 DMRs may result from different molecular mechanisms. | Different Mechanisms Cause Hypomethylation of Both H19 and KCNQ1OT1 Imprinted Differentially Methylated Regions in Two Cases of Silver–Russell Syndrome Spectrum
Silver–Russell syndrome is an imprinting disorder characterised by pre- and post-natal growth retardation and several heterogeneous molecular defects affecting different human genomic loci. In the majority of cases, the molecular defect is the loss of methylation (LOM) of the H19/IGF2 differentially methylated region (DMR, also known as IC1) at the telomeric domain of the 11p15.5 imprinted genes cluster, which causes the altered expression of the growth controlling genes, IGF2 and H19. Very rarely, the LOM also affects the KCNQ1OT1 DMR (also known as IC2) at the centromeric domain, resulting in an SRS phenotype by an unknown mechanism. In this study, we report on two cases with SRS features and a LOM of either IC1 and IC2. In one case, this rare and complex epimutation was secondary to a de novo mosaic in cis maternal duplication, involving the entire telomeric 11p15.5 domain and part of the centromeric domain but lacking CDKN1C. In the second case, neither the no 11p15.5 copy number variant nor the maternal-effect subcortical maternal complex (SCMC) variant were found to be associated with the epimutation, suggesting that it arose as a primary event. Our findings further add to the complexity of the molecular genetics of SRS and indicate how the LOM in both 11p15.5 DMRs may result from different molecular mechanisms.
Silver–Russell syndrome (SRS; OMIM #180860; also SRS2: #618905; SRS3: #616489; SRS4: #618907; SRS5: #618908. Estimated prevalence: 1:30,000–1:100,000) is a congenital disorder, characterised by intrauterine and post-natal growth retardation, relative macrocephaly at birth, feeding difficulties, a protruding forehead in early life, body asymmetry, and other less frequent features [1]. According to the Netchine-Harbison clinical score system (NH-CSS), clinical diagnosis is based on the presence of at least four out of the six most frequent features [1]. Recently, the definition of the Silver–Russell syndrome spectrum (SRSp) has been proposed to include all the cases with a clinical score < 4 but that still show clinical or molecular features of SRS [2]. The SRSp is caused by molecular changes affecting imprinted genes. The most frequently affected locus is at 11p15.5 and harbours two distinct imprinted domains overall extending for 1 Mb. A germline differentially methylated region (DMR) with the role of an imprinting centre (IC) regulating the monoallelic and parent-of-origin expression of the imprinted genes is present in each domain [3]. The telomeric domain encodes two genes with reciprocal imprinting: the insulin-like growth factor 2 (IGF2) gene is expressed from the paternal chromosome and encodes a foetal growth factor, and the H19 gene is expressed from the maternal chromosome and encodes a non-coding RNA with growth inhibitory activity [4]. These two genes are under the control of the H19-IGF2:IG (intergenic)-DMR (also known as IC1) that is normally methylated on the paternal allele. The centromeric domain encodes the growth inhibitor CDKN1C, which is transcribed from the maternal chromosome, and its in cis repressor KCNQ1OT1, which is transcribed from the paternal chromosome and is a non-coding RNA [4]. The imprinting of the centromeric domain is controlled by the KCNQ1OT1:TSS (transcription start site)-DMR (also known as IC2) that overlaps the promoter of KCNQ1OT1 and is normally methylated on the maternal allele. The most frequent molecular defects associated with the SRSp include a number of genetic and epigenetic alterations of the 11p15 imprinted gene cluster (accounting for 30–60% of cases), overall causing an increased expression of the growth inhibitory genes H19 and CDKN1C, and a decreased expression of the growth stimulatory gene IGF2. Maternal uniparental disomy of chromosome 7 (upd(7)mat) accounts for another 5–10% of cases [1]. Further alterations have been found at a lower frequency and affect either imprinted or non-imprinted genomic loci on different chromosomes (recently reviewed by Mackay and Temple [2]). For the molecular diagnosis of the SRSp, it is recommended to test the DNA methylation of both IC1 and IC2 first, and in case of a positive result, determine if the epigenetic abnormality is associated with any CNV or UPD to estimate the recurrence risk [1]. Although IC1 loss of methylation (LOM) occurs most frequently as an isolated primary epimutation, it can be associated with IC2 gain of methylation (GOM) as consequence of upd(11)mat [5] or maternally inherited duplications of the entire cluster in rarer cases [6]. Isolated IC2 GOM is even rarer and can be associated with either maternally inherited duplications [7] or paternally inherited deletions [8]. In a few cases, IC1 LOM has been found to be associated with the LOM of additional imprinted DMRs [9]. This latter condition is known as multi-locus imprinting disturbances (MLID) and can be associated with the clinical manifestation of features not usually present in SRS and dependent on the affected loci [10]. In some of these cases, IC1 LOM is detected together with IC2 LOM, which is a hallmark of the overgrowth-associated Beckwith–Wiedemann syndrome spectrum (BWSp) and shows the phenotypic features of either the SRSp or BWSp [11]. In some families, MLID has been associated with loss-of-function or hypomorphic variants of maternal-effect genes encoding protein components of the subcortical maternal complex (SCMC) [12]. Here we report on two cases with SRSp features and both IC1 LOM and IC2 LOM. In one patient, the epimutations are associated with a mosaic de novo 1.9 Mb duplication of maternal origin, involving the entire telomeric 11p15.5 domain and part of the centromeric domain but lacking CDKN1C. In the second patient, no clearly pathogenic genetic change is associated with the epimutations, which thus appear to be of primary origin. Our findings further add to the complexity of the molecular genetics of the SRSp and indicate how the LOM of both 11p15.5 DMRs may result from different molecular mechanisms.
Case 1. Proband 1 was born at term (37th w) from unrelated parents of Caucasian origin. The pregnancy was complicated by oligohydramnios. At birth, auxological parameters (weight 3.120 kg, 17th percentile/Z: −0.97; length 49 cm, 18th percentile/Z: −0.91; head circumference 35 cm, 53th percentile/Z: +0.08) were appropriate for gestational age. A spontaneously resolved mild atrial defect was reported at six months of age. A slight delay in language and acquisition of autonomous walking was observed in the first two years of life. Additionally, the face had a triangular shape with a prominent forehead. During childhood, the patient exhibited behavioural problems, relationship difficulties and learning difficulties. Since birth, the patient manifested gastrointestinal problems, including pyloric stenosis, intestinal occlusion, constipation, aversion to food and colic without any evidence of a precise organic cause. Between 5 and 8 years, height was <3rd percentile (Z: −2.8) and GH level was lower than normal (3.70 ng/mL after stimulation with arginine and 3.30 ng/mL after clonidine stimulation). At 8 years of age, the patient was treated with GH (Growth Hormone®) and during the first year of treatment he had already significantly gained height (32nd percentile/Z: −0.46). During the follow-up, a cystic formation of the pineal gland was demonstrated by brain MRI and a slight scoliosis was revealed by total spine X-ray. At our first clinical evaluation, the patient was 11 years and 11 months, his weight was 33.5 kg (13th percentile/Z: −1.13), height was 147.5 cm (44th percentile/Z: −0.15), cranial circumference was 54.3 cm (65th percentile/Z: +0.38), and BMI was 15.4 (11th percentile/Z: −1.20). We noted arched palate, slight prognathism, globose abdomen, several nevi on the back, clinodactyly of the 5th fingers, slight asymmetry of the lower limbs (<0.5 cm) and valgus hindfoot. Mild difficulties in motor coordination were also observed. A suspicion of SRS diagnosis was raised only in late childhood. However, according to the NH-CSS, only 2 of the 6 major criteria and some additional clinical features for SRS diagnosis were present. A molecular diagnosis of SRS with maternal 11p15 duplication was first indicated by the results of an SNP-array, which was proposed because of the neurodevelopmental delay. Case 2. Proband 2, son of unrelated parents of Caucasian origin, was born at term (39th w) after a pregnancy complicated by intrauterine growth retardation (IUGR). At birth, growth deficiency was evident (weight 2.170 kg, <3rd percentile/Z: −3.19; length 46 cm, 1th percentile/Z: −2.31; head circumference 32 cm, 1th percentile/Z: −2.51). During the first year of life, he was breastfed with frequent episodes of regurgitation. Delay of the anterior fontanel closure and growth retardation were observed. The patient had normal psychomotor development and a very sociable character. Since the 8th month of life, heterometry of the lower limbs was evident. At our first clinical evaluation the patient aged 2 years and 2 months, weight was 9 kg (<3th percentile/Z: −3.15), height was 80.5 cm (1th percentile/Z: −2.26), cranial circumference was 46 cm (3th percentile/Z: −1.94), and BMI was 13.9 (3°/Z: −1.8). Dolichocephaly (17.5 cm of biparietal diameter, 19 cm of anteroposterior diameter), a face of triangular shape with protruding forehead, nasal hypoplasia, clinodactyly of the V fingers, flat philtrum, asymmetry of the lower limbs, and the right limb 1.5 cm longer than left were also evident. Clinical assessment according to the NH-CSS showed a rating of 4.
Peripheral blood lymphocyte (PBL) genomic DNA of probands and their parents was isolated by a QIAsymphony automatic extractor (QIAGEN, Hilden, Germany).
Methylation-Specific Multiple Ligation-Dependent Probe Amplification (MS-MLPA) was performed on 50 ng of genomic PBL DNA by the commercially available assay, the SALSA MS-MLPA Probemix ME030-C3 or ME034-C1 (MRC-Holland, Amsterdam, The Netherlands), following manufacturer’s instructions. ABI 3500 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) was employed for the separation of the amplified products by capillary electrophoresis. Data were analysed using Coffalyser software (MRC-Holland, Amsterdam, The Netherlands). Bisulfite conversion and Pyrosequencing analysis was carried out as previously reported [13]. Briefly, 1.5 μg of genomic DNA was treated with sodium bisulfite by the EpiTect Bisulfite kit (Qiagen, Hilden, Germany, cat. n. 59104) following the manufacturer’s protocol. PyroMark PCR kit (Qiagen, Hilden, Germany, cat. n. 978705) was used to amplify 200 ng of converted DNA. Fifteen μL of PCR product was used for the quantitative analysis of DNA methylation by pyrosequencing on a Pyromark Q48 Autoprep system with the PyroMark Q48 Adv. CpG Reagents (Qiagen, Hilden, Germany cat. n. 974022) and PyroMark Q48 Magnetic Beads (Qiagen, Hilden, Germany cat. n. 974203). Results were analysed by the Pyromark Q48 Autoprep software. The primer sequences have been previously reported [13]. Methylome array was performed on bisulphite converted PBL DNA of proband 2. Data were analysed using R version 4.1.0. β values were extracted from “idat” files by using the “champ.load” module of the “ChAMP” R package (v.2.22.0), with quality control options set as default and array type as “EPIC.” To adjust the β -values of type 2 probes, we applied BMIQ normalization with the default options for array type as “EPIC.” The coordinates of the imprinted DMRs were downloaded from http://www.humanimprints.net/ (accessed on 20 December 2021). Methylation profile was calculated as average of the methylation levels of their respective CpGs. Methylation levels of the patient were compared with 4 age-matched controls; a value deviating ± 3 standard deviation from the mean of the controls was considered as an aberrant methylation change. The raw and processed files are available on request.
Single nucleotide polymorphism-array (SNP-array) analysis was performed on DNA of proband 1 and his parents using Infinium CytoSNP-850 K BeadChip (Illumina, San Diego, CA, USA) and in accordance with the manufacturer’s instructions. Array scanning data were generated by iScan system (Illumina, San Diego, CA, USA) and the results were analysed by Bluefuse Multi software (v 4.4).
Fluorescence in situ hybridization (FISH) analysis was performed to provide structural information on the microduplication. Locus-specific FISH analysis was performed on metaphases and nuclei obtained from PHA-stimulated lymphocytes, by means of a custom oligonucleotide probe (SureDesign, Agilent, Santa Clara, CA, USA), specifically designed within the duplicated region (11p15.5).
Molecular testing for SRS was performed by MS-MLPA (ME030 assay) on the PBL DNA of the two patients who had received a clinical diagnosis of SRS (proband 2) or were suspected to be on the SRSp (proband 1; see Figure 1). The methylation analysis revealed the IC1 and IC2 LOM in both the probands, but it was less severe in proband 1 (methylation level 39.5% of both ICs) than in proband 2 (30%). The methylation defect was confirmed by a sodium bisulphite treatment and pyrosequencing in the probands and was excluded in their parents (Figure S1). The MLPA analysis also revealed a microduplication of 11p15.5, including both H19 and KCNQ1OT1 in proband 1. The last exons of KCNQ1, located downstream to IC2, as well as the CDKN1C gene were not included in the duplication. The copy number value (<1.5) suggested the presence of the duplication in the mosaic form. No CNV at 11p15.5 was detected in proband 2 (Figure 1). The microduplication of proband 1 was further characterized by an SNP-array analysis, which confirmed the presence of a de novo 11p duplication in the mosaic form, with an extension of about 1.9 Mb, involving the entire telomeric SRS/BWS domain and only part of the centromeric domain (Figure 2A). The breakpoints were mapped at positions 795,147 and 2,712,286 of chr 11p (GRCh37). The lack of SNP-array probes within the IC2 region did not allow for the detection of IC2 CNVs. However, all four MS-MLPA probes for the IC2 CNV analysis revealed the duplication, demonstrating that at least two-thirds of the DMR and at least 200 bp centromeric to the transcription start site of KCNQ1OT1 are included in the duplication. Furthermore, the analysis of the SNP genotypes of the duplicated region in the trio demonstrated that the duplication was of maternal origin (Figure 2B). The metaphase FISH analysis on the proband lymphoblasts demonstrated the hybridization of 11p15-specific probes only at the telomeric region of chromosome 11 and a signal of stronger intensity on one homologue in about half of the analysed cells, indicating that the duplication was in tandem and present in the mosaic form (Figure 2C). As no CNV was detected at 11p15.5, further analyses were performed to investigate if the epigenetic defect was extended to further imprinted loci in proband 2. The multi-locus MS-MLPA (ME034) did not reveal any epigenetic defects or CNV at additional disease-associated imprinted loci (Figure S2A). The methylation analysis of 39 imprinted DMRs by an Illumina Infinium EPIC methylation array confirmed the 11p15.5 LOM and demonstrated a further slight LOM at NAP1L5:TSS-DMR on chr 4q22 and a slight GOM affecting the INPP5F:Int2-DMR on chr 10q26 (Figure S2B). Overall, the MS-MLPA and methylome results showed that the 11p15.5 ICs epimutations of proband 2 are associated with an MLID profile. Whole-exome sequencing was performed on the DNA of proband 2’s mother, but no clearly pathogenic variant in the SCMC genes was identified.
Among the known imprinting disorders, SRS is probably the one with the most heterogeneous molecular genetics, a fact that makes molecular diagnosis very challenging. Accordingly, five different OMIM ID entries have been associated with SRS so far (https://omim.org (accessed on 2 March 2022); chromosome 11p15.5 (IC1: SRS1, #180,860; IGF2: SRS3, #616,489), 7p13-q32 (SRS2, #618,905), 8q12.1 (SRS4, PLAG1, #618,907), and 12q14 (SRS5, HMGA2, #618,908)). The majority of the cases belong to the first subgroup which is further characterized by different molecular mechanisms underlying IC1 LOM. In this study, we report two cases that further add to the molecular complexity of SRS1. Both patients showed a LOM of either IC1 and IC2, but this complex epimutation was associated with a de novo mosaic in cis maternal duplication in the former case, but no 11p15.5 CNV or maternal-effect SCMC variant in the latter case. Case 1. Proband 1 represents a peculiar case both due to his clinical features and molecular defect. Concerning the clinical diagnosis, the proband does not fulfil the clinical criteria of SRS according to NH-CSS, as only two (post-natal growth retardation and protruding forehead) out of six main features are reported. Although molecular testing demonstrated IC1 LOM, this case does not show the typical defects of SRS, because the maternal duplication is not extended to the whole centromeric 11p15.5 domain and particularly does not include the CDKN1C gene. Nevertheless, several additional features of SRS (slight asymmetry, fifth finger clinodactyly, triangular face, scoliosis, speech and motor delay) are present, and the duplication includes the telomeric domain with the H19 gene. According to a definition recently proposed by Mackay and Temple, this case might be possibly classified within the Silver–Russell syndrome spectrum (SRSp) [2]. Although the duplication has a maternal origin, IC2 is affected by the LOM instead of the expected GOM due to its maternal methylation. This discrepancy could result from a position effect of nearby loci on the duplicated IC2, which is located at the end of the duplicated region (Figure 3). In three previously reported cases, partial duplications of the centromeric domain were associated with IC2 LOM but with the BWS phenotype when maternally inherited [14,15]. Indeed, IC2 LOM is expected to cause the activation of KCNQ1OT1, the repression of CDKN1C, and most likely, overgrowth. Our patient shows normal growth parameters at birth and a mild SRS phenotype in infancy, which may possibly result from a compensatory effect between the growth stimulation deriving from IC2 LOM and the growth inhibition caused by H19 duplication (Figure 3). Alternatively, it is possible that the duplicated KCNQ1OT1 is not expressed because it lacks part of its promoter or because KCNQ1OT1 is too far away (>2–4 Mb) from CDKN1C to exert its repressive action in cis because it is inserted in the telomeric breakpoint. Maternal 11p15.5 duplications associated with SRS are generally germline and affect the entire imprinted cluster or are restricted to its centromeric domain [1,7,14], supporting a role of CDKN1C in SRSp pathogenesis. Although H19’s function as growth inhibitor is indicated by several mouse studies [16,17], the role of the telomeric domain in SRS is less clear. A case of the maternal duplication of the entire telomeric domain was associated with a normal phenotype [18]. Three further cases with the partial duplication of the telomeric domain, involving H19 but not IGF2, displayed growth retardation, but two of them had additional cytogenetic anomalies that might also explain this phenotype [14,15,19]. Finally, a somatic maternal 11p15 duplication has been identified on the smaller side of the face of a patient with body asymmetry [20]. The duplication of our patient is mosaic, maternal, includes the entire telomeric domain and the centromeric domain but not CDKN1C, and is not associated with any further cytogenetic anomalies. This finding supports a role of the duplicated H19 in SRSp pathogenesis, although the involvement of further genes located in the duplicated region cannot be completely ruled out. By querying the Decipher database, we have found that about fifty genes in addition to the 11p15.5 imprinted gene cluster are included in the duplication of proband 1 (https://www.deciphergenomics.org/search/patients/results?q=grch37%3A11%3A795147-1941891, accessed on 15 September 2022). Twenty-six cases are reported to be carriers of duplications overlapping this region. Of the very heterogeneous clinical features of these individuals, a few are present in proband 1, such as mild developmental delay, short stature, and clinodactyly. However, in most of these cases, the duplication also involves the imprinted gene cluster, making a correct genotype–phenotype correlation difficult. On the other hand, it cannot be excluded that genes outside the 11p15.5 imprinted gene cluster may contribute to the atypical clinical phenotype of proband 1. Case 2. IC1 LOM and IC2 LOM are hallmarks of SRS and BWS, respectively. Nevertheless, both these epimutations are associated with severe growth retardation in proband 2. IC1 LOM+IC2 LOM has previously been reported in several cases, whose phenotype was either SRS or BWS (Table 1). As in proband 2, the epigenetic defect is always partial, supporting the hypothesis of errors in imprinting maintenance arising post-zygotically. The resulting epigenetic and gene expression mosaicism probably explain the divergent clinical features as well as the frequent body asymmetry of the affected individuals [11,12]. Most of the IC1 LOM+IC2 LOM patients show the hypomethylation of additional DMRs and have been classified as MLID cases (Table 1). Although further studies are needed to clarify genotype–epigenotype correlations, the current idea to explain the clinical outcome of the MLID patients is the epidominance hypothesis, which is based on the mosaic form of the multiple methylation changes in BWS and SRS [11]. According to this hypothesis, the main clinical presentation of the patient is caused by the imprinted locus that is mostly affected, while the other affected loci may possibly contribute to atypical features. For example, the LOM of the GNAS locus has been found associated to 11p15.5 IC2 LOM in MLID patients with BWS and pseudohypoparathyroidism 1B [21] or hypocalcemia [22]. In SRS-MLID, the most affected DMRs other than IC1 are MEST:alt-TSS-DMR and GRB10:alt-TSS-DMR [12] (Table 1), although we did not find abnormal methylation at these loci in our patient. Some cases of BWS and SRS with MLID have been associated with maternal variants of the SCMC genes [9,12,22,23,24,25,26,27], but the whole-exome sequencing did not identify any such variant in our case. | true | true | true |
|
PMC9602392 | Cui Cao,Zhongfu Wang,Guiping Gong,Wenqi Huang,Linjuan Huang,Shuang Song,Beiwei Zhu | Effects of Lycium barbarum Polysaccharides on Immunity and Metabolic Syndrome Associated with the Modulation of Gut Microbiota: A Review | 12-10-2022 | Lycium barbarum polysaccharides,structural characteristics,gut microbiota,immunity,metabolic syndrome | Lycium barbarum polysaccharides (LBPs) have attracted increasing attention due to their multiple pharmacological activities and physiological functions. Recently, both in vitro and in vivo studies have demonstrated that the biological effects of dietary LBPs are related to the regulation of gut microbiota. Supplementation with LBPs could modulate the composition of microbial communities, and simultaneously influence the levels of active metabolites, thus exerting their beneficial effects on host health. Interestingly, LBPs with diverse chemical structures may enrich or reduce certain specific intestinal microbes. The present review summarizes the extraction, purification, and structural types of LBPs and the regulation effects of LBPs on the gut microbiome and their derived metabolites. Furthermore, the health promoting effects of LBPs on host bidirectional immunity (e.g., immune enhancement and immune inflammation suppression) and metabolic syndrome (e.g., obesity, type 2 diabetes, and nonalcoholic fatty liver disease) by targeting gut microbiota are also discussed based on their structural types. The contents presented in this review might help to better understand the health benefits of LBPs targeting gut microbiota and provide a scientific basis to further clarify the structure–function relationship of LBPs. | Effects of Lycium barbarum Polysaccharides on Immunity and Metabolic Syndrome Associated with the Modulation of Gut Microbiota: A Review
Lycium barbarum polysaccharides (LBPs) have attracted increasing attention due to their multiple pharmacological activities and physiological functions. Recently, both in vitro and in vivo studies have demonstrated that the biological effects of dietary LBPs are related to the regulation of gut microbiota. Supplementation with LBPs could modulate the composition of microbial communities, and simultaneously influence the levels of active metabolites, thus exerting their beneficial effects on host health. Interestingly, LBPs with diverse chemical structures may enrich or reduce certain specific intestinal microbes. The present review summarizes the extraction, purification, and structural types of LBPs and the regulation effects of LBPs on the gut microbiome and their derived metabolites. Furthermore, the health promoting effects of LBPs on host bidirectional immunity (e.g., immune enhancement and immune inflammation suppression) and metabolic syndrome (e.g., obesity, type 2 diabetes, and nonalcoholic fatty liver disease) by targeting gut microbiota are also discussed based on their structural types. The contents presented in this review might help to better understand the health benefits of LBPs targeting gut microbiota and provide a scientific basis to further clarify the structure–function relationship of LBPs.
The human gut microbiota is a complex and abundant community composed of up to 1014 microorganisms with about 1150 species [1]. The community is dominated by Firmicutes and Bacteroidetes, which account for more than 80–90%, and then followed by Proteobacteria, Actinobacteria, Verrucomicrobia, Fusobacteria, Cyanobacteria, and Spirochaetes as minor components [2]. The gut microbiota is regarded as a neglected human organ to some extent in the human–microbe superorganism [3]. Furthermore, the dysbiosis of gut microbiota not only affects the host physiological functions (e.g., nutrient digestion, absorption, and metabolism), but triggers diseases (e.g., immune dysregulation responses and metabolic syndrome) [4,5,6]. Therefore, the balance of gut microbiota, including microbial diversity, richness, composition, and functionality, is critical for the health of the host. Numerous studies have demonstrated that several factors, such as genetics, antibiotics, age, and diet, can influence the gut microbiome [6,7]. Among these factors, a short-term diet can lead to significant microbial changes. More importantly, non-digestible polysaccharides can be degraded and utilized by gut microbiota instead of the host, which encode the carbohydrate active enzymes (CAZymes), such as glycoside hydrolases (GHs), polysaccharide lyases (PLs), glycosyltransferases (GTs) and carbohydrate esterases (CEs), thereby improving beneficial metabolites (e.g., SCFAs) [8,9]. Lycium barbarum, also named Goji berry, Gouqizi, and wolfberry, is a perennial shrubbery of Solanaceae that is widely cultivated in China, Japan, Korea, North America, and Europe [10]. Currently, China is the largest supplier in the world, and a majority of L. barbarum fruits are distributed in the northwest regions of China, such as Ningxia, Xinjiang, Tibet, Inner Mongolia, Qinghai, and Gansu [11,12]. Notably, L. barbarum fruits from Ningxia region are the only species included in the Pharmacopoeia of the People’s Republic of China for many years due to their excellent quality [13]. Various bioactive constituents have been isolated and identified from L. barbarum fruits, including polysaccharides, carotenoids, vitamins, flavonoids, alkaloids, anthraquinones, anthocyanins, and organic acids. Among them, the polysaccharides, accounting for 5–8% of dried fruits, have been recognized one of the principal active components [10]. In recent decades, a great deal of research has now confirmed that L. barbarum polysaccharides (LBPs) have various biological functions, such as immunoregulation, anti-inflammation, anti-tumor activities, hypoglycemic/lipidemic activities, and retinal protection [14,15,16,17,18,19]. LBPs mainly include arabinogalactans, acidic heteropolysaccharides, glucans, and other polysaccharides [20,21,22,23,24]. Increasing evidence suggests that the molecular weight, monosaccharide composition, and glycosidic linkage of LBPs could influence their bioactivities, although the structure–activity relationship of polysaccharides is not yet clear. Therefore, elucidating the structures of LBPs would be beneficial to understand the mechanisms of their health effects and further develop their industrial application. However, many studies have shown that most LBPs are resistant to human digestive enzymes and can almost entirely reach the colon where they are digested and metabolized by gut microbiota, indicating that gut microbiota plays a crucial role in the beneficial effects of LBPs [25,26]. Currently, although the extraction, purification, structural characterization, and functional activities of LBPs have been summarized and reviewed [27,28,29], few reviews have discussed their structural types and summarized the modulation of LBPs on gut microbiota and the role of gut microbiota in the health effects of LBPs, as well as their potential mechanism based on their structural types. This review mainly summarizes the modulation of LBPs on gut microbes and related metabolites. Furthermore, the protective effects of LBPs mediated by gut microbiota on immunoregulation (e.g., immunopotentiation and anti-inflammation), metabolic disease (e.g., obesity, type 2 diabetes, and nonalcoholic fatty liver disease), and other diseases (e.g., asthma and emotional impairment) have been summarized and discussed in order to better understand the health benefits of LBPs targeting gut microbiota in the present review. In addition, the current issues and future prospects for the relationship between the structure and function of LBPs are also discussed.
The elucidation of precise structures of LBPs is the prerequisite to unraveling the relationships between structures and functions. Numerous studies have demonstrated that the biological activities of LBPs are principally related to their primary and advanced structures [10,28]. Actually, the current studies mainly focus on the primary structures of LBPs due to the limitations of techniques and analysis. The primary structure characterization of LBPs covers molecular weight, types and ratios of monosaccharides, positions of glycosidic linkages, anomeric carbon configuration, and branched chains, which influence their biological activities to varying degrees [18,24]. Herein, the research progress on the extraction, purification, and structure of LBPs were summarized below.
The isolation principle of LBPs is to keep the properties of polysaccharides unaltered during the procedure of extraction and purification. Based on this principle, several extraction methods for crude LBPs have been developed, which include cold or hot water extraction, microwave-assisted extraction, enzyme-assisted extraction, ultrasonic-assisted extraction, and supercritical fluid extraction [10,27]. Indeed, water extraction is the most commonly used method to obtain crude LBPs due to its convenient operation and high yield [27,30]. For example, high molecular weight polysaccharides were obtained from dried wolfberries using cold water extraction in a yield of 2–3%, however, the yields of the polysaccharides could be further improved by prolonged high-temperature extraction or enzymatic treatment [30]. Furthermore, it demonstrated that a ratio of water to raw material 31.2, temperature 100 °C, time 5.5 h, and number of extraction 5 were the optimal extraction conditions to obtain LBPs using the Box–Behnken statistical design (predicted yield 23.13%), which was verified by validation experiments (real yield 22.56 ± 1.67%) [31]. Given the excellent solubility of LBPs in water, several scholars have argued that the increased LBPs contain more pectic, cellulose, and hemicellulosic polysaccharides by extended treatments, such as high temperature, enzymatic treatment, and microwave-assisted treatments [31,32]. Generally, the water-soluble extracts using the above extraction methods contain many impurities, such as inorganic salts, pigments, monosaccharides, oligosaccharides, and proteins, which interfere with the structure determination of LBPs. Therefore, effective measures have to be adopted to further purify the above crude LBPs. Hydrogen peroxide, as a chemical reagent, is widely applied in depigmentation and the Sevag method is frequently applied in deproteinization for their simple procedures [33]. Subsequently, the methods for LBP purification can be performed by membrane separation (e.g., ultrafiltration and microfiltration), column chromatography (e.g., gel filtration chromatography, ion-exchange chromatography, affinity chromatography, and cellulose column chromatography), and chemical precipitation (e.g., fractional precipitation with ethanol) alone or in combination [27,33]. Of note, column chromatography is most commonly used in these methods [27]. As we previously reported, five arabinogalactan fractions (LBP1~5) from crude LBPs (extracted by water at room temperature) were separated by DEAE-cellulose chromatography [34]. Afterwards, LbGp1 with a molecular weight of 49.1 kDa was isolated and purified from LBP1 by Sepharedax G-100 column chromatography in yields of 0.018% [22]. Similarly, another five fractions (LRP1, LRP2, LRP3, LRP4, and LRP5) were also isolated from crude L. ruthenicum polysaccharides (extraction by 70 °C water) on DEAE-Cellulose-52 anion-exchange column followed by gradient elution in our previous studies [35]. Subsequently, LRGP1 (Mw 56.2 kDa) and LRGP3 (Mw 75.6 kDa) were further purified on Sephadex G-100 column in yields of 0.003% and 0.008%, respectively [35,36]. Moreover, LBP3b (Mw 5 kDa) was purified from crude LBPs extracted with hot water (60 °C) using DEAE-cellulose column and Sephadex G-150 column, which was identified as glucan [24]. In addition, a novel arabinogalactan LBP1A1-1 (Mw 45 kDa) was purified from L. barbarum on DEAE Sepharose Fast Flow column and Sephacryl S-200 HR column in yields of 0.1% [37]. These studies have indicated that the polysaccharide fractions purified by column chromatography are difficult to investigate for the activities in vivo, as well as the structure–function relationship due to low yield and complex operation. Then, we developed fractional precipitation with 30%, 50%, and 70% (V/V) ethanol to purify arabinogalactan in yields of 0.38%, which was simpler and more efficient than column chromatography [17].
To date, LBPs have been identified as glycoconjugates that mainly consist of five major structural elements: arabinogalactan, pectin polysaccharide, glucan, xylan, and other heteropolysaccharides [21,22,23,24]. Their hypothetical structure features, such as monosaccharide composition, repeat unit, and molecular weight, were summarized in Table 1. Additionally, the molecular weight of LBPs is highly subject to the origin, cultivar, and extraction method, ranging from 5 kDa to 2300 kDa [10,24,38].
Structural characterization of L. barbarum arabinogalactan-protein has been investigated by multiple research groups, and it has been demonstrated that there are a large number of →3,6)-Galp-(1→ residues based on the methylation analysis. The current controversies about its structure are as follows: (1) L. barbarum arabinogalactan is composed of →6)-β-Galp-(1→ as the backbone, and large amounts of α/β-Araf as branch chains which substituted at C-3 [22,41,48] (Figure 1A); (2) it is a highly branched polysaccharide with a backbone of →3)-β-Galp-(1→ substituted at C-6 with Araf [40,45] (Figure 1B); (3) the fraction possesses both β-(1→6)-linked Galp and β-(1→3)-linked Galp as the backbones with partial substitution at the C-3 site and C-6 site, respectively [37,42] (Figure 1C). The backbone structure of arabinogalactan in LBPs may be different due to diverse origin and various isolation methods. As mentioned above, a combination of ion exchange column and gel filtration column chromatography is commonly employed for the purification of arabinogalactan fraction from L. barbarum glycoconjugates; however, it is not suitable for large-scale preparation of arabinogalactan due to complex operation, time-consuming processes, and low yield. Recently, our research team revisited the structure of L. barbarum arabinogalactan using a set of chemical methods and analytical techniques, including partial acid hydrolysis, methylation analysis, alkaline degradation, monosaccharide composition analysis, 1H and 13C spectroscopy, and ESI-MSn [39] on the basis of the ethanol precipitation method reported [17]. And the results indicated that it was a highly branched polysaccharide with a backbone of →6)-β-Galp-(1→ and branched chains of →3)-β-Glap (1→, →3)-α-Araf-(1→ and →5)-β-Araf-(1→ substituted at the C3 position, which had an average of 9 branches per 10 sugar backbone units. Additionally, the anti-aging activity of L. barbarum arabinogalactan was significantly higher than the backbone fraction (Gal percentage = 91%) obtained by partial acid hydrolysis (0.02 M H2SO4), indicating that the anti-aging activity was closely relevant to the arabinose branched chains. These results implied that the biological activities of LBPs were considerably influenced by their structures, especially branched chains and spatial configuration [39].
Pectins, as a cell wall component of plants, are unique polysaccharides comprising predominantly uronic acids, such as glucuronic acid (GlcA) and galacturonic acid (GalA) [57]. The polysaccharides extracted from L. barbarum fruits also contain pectins (Figure 1D). There are mainly three typical structures in pectins: homogalacturonan (HG), rhamnogalacturonan-I (RG-I), and rhamnogalacturonan-II (RG-II) [58,59]. A typical pectic polysaccharide (p-LBP) with a backbone of →4-α-GalpA-(1→ (HG) and a partial region of →4-α-GalpA-(1→ and →2-α-Rhap-(1 → (RG-I) was isolated and purified using a series of column chromatographies (e.g., macroporous resin S-8, DEAE column and Sephacryl S400 gel permeation) and analytical techniques (e.g., 1H and 13C spectroscopy) [23]. Another acidic polysaccharide (LBP3a) was also separated from the crude extraction by DEAE-cellulose chromatography, which was identified as HG-type pectin with a backbone of →4)-α-D-GalpA(1→ [53]. HG-type pectin was found in the above studies, perhaps due to the same extraction methods (e.g., hot water) and original place. Besides, the polysaccharides from L. barbarum insoluble cell wall material (CWM) dissolved in the CDTA and Na2CO3 solutions contained 76.3% and 51.9% uronic acid, respectively. Notably, the fraction extracted by CWM-Na2CO3 may be RG-type pectin, which was supported by the increased level of rhamnose (Rha) [46]. Additionally, one homogeneous polysaccharide (LBP-1, Mw 2250 kDa) was purified from crude LBPs using DEAE column, whose structure was identified as pectin with a backbone of α-(1→5)-l-Ara and α-(1→4)-d-GalA, and branched chains of →3)-Man-(1→, →6)-Man-(1→, and T-Man-1(→ [38].
Glucans widely exist in the cell walls of various plants and fungi, and there is a small amount in L. barbarum fruits, despite the diversity in conformation and linkages [60]. For instance, LBP1a-1 (Mw 115 kDa) and LBP1a-2 (Mw 94 kDa) were obtained from crude LBPs using DEAE-cellulose and Sephacryl S-400 HR column chromatography, which was identified as glucan with a backbone of →6)-α-d-Glcp (1→ [53]. Moreover, a homogenous polysaccharide with a molecular weight of 4.9 kDa was separated from crude LBPs by the DEAE-cellulose column in combination with Sephadex G-150 column and then identified as a β-glucan by monosaccharide composition and 1H/13C NMR analysis [24]. In addition, an α-(1→4) (1→6) glucan (LBPC4) was isolated and purified from crude LBPs using DEAE-cellulose column and Sephadex G-50 column [55].
Xylans are the primary hemicellulose component in plant cells, which are mainly found in hardwood (15–30%), softwoods (7–10%), and annual plants (up to 30%) [61]. Additionally, 4 M KOH-soluble fraction isolated from L. barbarum insoluble cell wall material was a xylan instead of xyloglucan, which was supported by the fact that the xylose content was twice that of the glucose [46]. In addition, a β-(1→4) (1→6)-linked heteropolysaccharide (LBPC2) was separated from crude LBPs using DEAE-cellulose column and Sephadex G-50 column [55]. Interestingly, it was composed of only Xyl, Rha, and Man in a molar ratio of 8.8:2.3:1.0, so LBPC2 was supposed to be a xylan, which needs further confirmation.
Apart from the above four types, the structural elements of LBPs have been identified as other types from their monosaccharide composition in a few studies. For example, LBP-IV, which is mainly composed of Glc, Ara, and Xyl in a molar ratio of 7.54:3.82:3.44, was separated from crude LBPs on the DEAE-Sephadex A-25 column [56]. Another polysaccharide was isolated from crude LBPs with a macroporous resin S-8 column, which primarily comprised Glc, Man, and Rha in molar ratios of 6.52:2.17:0.81 [26]. These results indicate that LBPs contain other heteropolysaccharides in addition to arabinogalactan, pectin, glucan, and xylan; however, the structures need to be further identified and confirmed.
Generally, the polysaccharide chains of LBPs are primarily digested and utilized by gut microbes instead of the host. More concretely, they are hydrolyzed by microbial CAZymes (e.g., GHs and PLs) which are absent in the human genome. For example, the transfer rates of fluorescein isothiocyanate (FITC)-labeled LBP (arabinogalactan-pectin complexes) from basolateral to apical side and vice versa in Caco2 cell monolayer model were 0.98 and 0.92%, respectively, indicating that the transmembrane transport of LBP was extremely limited [62]. Furthermore, LBPs (arabinogalactan type) was not degraded under simulated saliva, gastric, and intestinal conditions, however, it could be utilized and metabolized by gut microbiota based on the consumption of total carbohydrates and promotion of SCFAs after fermentation in vitro [25]. Meanwhile, the above LBPs significantly improved the levels of Bacteroidetes (e.g., Bacteroides and Prevotella), Firmicutes (e.g., Lactococcus and Faecalibacterium), and Actinobacteria (e.g., Bifidobacterium), perhaps due to the carbohydrate degrading systems of Bacteroidetes (starch utilization system-like systems), Firmicutes, and Actinobacteria (ATP-binding cassette transporters), which implied that LBPs were degraded and utilized by gut microbes in a cooperative manner [8,25]. In addition, an LBP, comprising Glc, Man, Rha, Gal, Ara, Xyl in molar ratios of 6.52:2.17:0.81:0.23:0.18:0.07, markedly promoted the proliferation of the probiotic Bifidobacterium and Lactobacillus strains by improving the carbon and energy metabolism [26]. Notably, the activity of carbohydrate metabolism enzymes was significantly enhanced by LBP, especially β-galactosidase and lactate dehydrogenase [26]. Actually, microbial culture is an effective method to know and understand the degradation and utilization of LBP by gut microbiota in human health, however, almost none of the existing studies have been applied it to investigate the degradation and utilization mechanism by gut microbiota, perhaps due to the following reasons: (i) the complex structure of LBPs with high branches [27]; (ii) more than 80% of intestinal microbial species are uncultured in vitro [63]; (iii) the specific glycan preference of microbial species [64]; (iv) the cooperation among microbial species [65]. Currently, our research team has explored the microbial degradation of LBPs in pure culture, and two Bacteroides species that effectively utilized arabinogalactan from L. barbarum have been screened (unpublished data).
The dynamic balance of gut microbiota, including the microbial composition and its relative abundance, plays a key role in host intestinal homeostasis [2,66,67]. Numerous studies have demonstrated that the relative abundance of Firmicutes (~64%), Bacteroidetes (~23%), and Proteobacteria (~4.5%) account for over 90% at the phylum level, and any alteration in the microbial proportion tends to the intestinal immune dysregulation and even pathological changes [68,69,70]. Among them, Proteobacteria contains many well-known pathogens such as Shigellosis, Vibrio, Salmonella typhimurium, Escherichia coli, Staphylococcus aureus, Helicobacter pylori, and Pseudomonas aeruginosa. LBPs (without chemical characterization) remarkably inhibited the proliferation of pathogenic E. coli, S. typhimurium, and S. aureus in vitro [71,72]. Furthermore, sulfated LBPs with sulfation degrees of 1.5–2.0 could significantly improve antiviral (Newcastle disease virus) activity [73]. Additionally, LBPs with concentrations of 8–20 mg mL−1 not only suppressed the growth of E. coli in vitro, but reduced cecal E. coli in tumor mice [74]. Anomalous expansion of Proteobacteria (belonging to Gram-negative bacteria) is the microbial signature of dysbiosis in gut microbiota, and its level can be at least three times higher in inflammation and cancer (14.9%) than that in healthy humans (4.5%) [68]. What is more, Gram-negative bacteria produced more than twice as many pro-inflammatory cytokines IL-6 and IL-8 from human monocytes compared to Gram-positive bacteria [75]. Furthermore, compared to other cell wall constituents of bacteria such as peptidoglycan and teichoic acid, lipopolysaccharide (LPS) is the most efficient endotoxin isolated from bacteria cell walls to induce pro-inflammatory cytokines, which can be recognized by pattern recognition receptors (e.g., TLR4) [76]. The binding of LPS to TLR4 activates the MAPK/NF-κB signaling pathways, and culminates in the generation of pro-inflammatory cytokines (e.g., TNF-α), which is possibly responsible for the intestinal dyshomeostasis caused by pathogenic Proteobacteria, thereby exacerbating inflammation [75,77]. Our previous study reported that LBP-3 (arabinogalactan type) could significantly decrease the abundance of Proteobacteria in DSS-induced colitis mice, especially the pro-inflammatory Enterobacteriaceae, and inhibited the activation of TLR4-MAPK/NF-κB signaling pathways, thereby reducing levels of pro-inflammatory cytokines such as IL-1β and TNF-α [78]. Similarly, LBP (glucan type) remarkably downregulated the level of Proteobacteria, and reduced the LPS/TLR4/NF-κB signaling path in high-fat diet (HFD)-induced nonalcoholic fatty liver disease (NAFLD) in Sprague–Dawley (SD) rats [79]. In addition, supplementation with LBP-W (arabinogalactan type) markedly reversed the relative abundance of Proteobacteria induced by a HFD, turning it toward the normal level [41].
The promotion effect of LBPs on microbial richness and diversity is partially attributed to their probiotic function. An appropriate abundance of probiotics such as Bifidobacterium and Lactobacillus contributes to the maintenance of intestinal epithelial barrier function and the modulation of immune homeostasis by competitive inhibition of pathogens and generation of antimicrobial compounds (e.g., bacteriocins, lactate, and acetate), thereby reducing the inflammation triggered by harmful intestinal bacteria [80]. It has been demonstrated that LBP (without chemical characterization) with concentrations of 12–20 mg mL−1 significantly promoted the proliferation of Lactobacillus in vitro [74]. Similarly, LBPs mainly composed of Glc, Man, and Rha in molar ratios of 6.52:2.17:0.81 could pronouncedly improve the growth of B. bifidum, B. infantis, B. longum, B. animalis, L. acidophilus, and L. plantarum in vitro [26]. Furthermore, the same type of LBP as the above [26] supported the growth of L. acidophilus and B. longum with a maximum of 8.23 ± 0.30 (log10 CFU/mL) and 6.34 ± 0.11 (log10 CFU/mL), respectively, in de Man Rogosa Sharpe (MRS) broth; and administration of LBPs to normal mice also markedly improved the relative abundance of probiotic Lactobacillus, and enriched sIgA in the colon, thus enhancing the innate immunity [81]. In addition, supplementation with arabinogalactan-type LBP-W (50 mg kg−1 d−1) not only improved the diversity of gut microbiota but significantly increased the relative abundance of Lactobacillus in normal mice and HFD-induced obese mice [41]. More importantly, Ara, Gal, arabino-oligosaccharide, and galacto-oligosaccharide (GOS), as prebiotics, have been indicated to have the proliferative capacity of Bifidobacterium, which probably explains why L. barbarum arabinogalactans have the prebiotic effect [82,83,84]. Of note, Bifidobacterium and Lactobacillus (e.g., live combined Bifidobacterium and Lactobacillus tablets) have been widely used in the clinical treatment of pediatric gastrointestinal diseases (e.g., diarrhea) [85,86]. These research findings suggest LBP is a good potential prebiotic which can boost beneficial bacteria levels, modulate the intestinal microbiota structure, and regulate the intestinal homeostasis of the host.
Apart from the above enteric pathogens and probiotics, some commensal microbiota that are well-known glycan utilizers, such as Akkermansia, Prevotella, Bacteroide, Ruminococcaceae, Prevotellaceae, and Bacteroidaceae, can also be enriched by LBPs. These polysaccharide utilizers contain various GHs and PLs which are responsible for the degradation of polysaccharides [7]. We found that the various types of LBPs in similar experimental models could increase the level of Bacteroides, such as in the fermentation of arabinogalactan-type [25] and pectin-type [87] LBPs by the human gut microbiota in vitro. Similarly, Akkermansia, hailed as an emerging “second generation” probiotic, was also markedly elevated in Kunming mice with a normal diet [81] and in C57BL/6J mice with a normal diet [88] by different LBPs. Furthermore, the relative abundance of Ruminococcaceae, known as secondary bile acids-producing bacteria [89], was also significantly improved by LBPs in normal mice [90] and DSS-induced colitis mice [78]. Unlike the above findings, the levels of SCFA-generating bacteria were altered to various degrees. For example, arabinogalactan-type LBPs could significantly augment the abundance of Bacteroidaceae, Lachnospiraceae, and Ruminococcaceae in cyclophosphamide (CTX)-induced immunocompromised BALB/c mice [91]. In comparison, arabinogalactan-pectin complex WBPPS not only upregulated the levels of Bacteroidaceae and Ruminococcaceae, but also downregulated Rikenellaceae, Marinifilaceae, and Alistipes in CTX-induced mice [92]. Furthermore, supplementation with LBP (without chemical characterization) decreased the relative abundance of A. muciniphila, Allobaculum stercoricanis, Citrobacter, Tannerella, Spirochaeta, and Parasutterella excrementihominis in normal C57BL/6J mice fed with a standard diet [88]. Hence, the effects of LBPs on gut microbiota are complicated, perhaps depending on the types of LBPs and animal models. In summary, the beneficial effects of LBPs on the host health may be attributed to the enrichment of probiotics, the decrease of pathogens, and the stabilization of symbiotic bacteria, i.e., its capacity for balancing microbial structure.
Small molecule metabolites that are generated as intermediate or final products by gut microbiota play a crucial role in the interaction between gut microbiota and the host, which contributes to the modulation of intestinal and systemic immunity. Given that the gut microbiota is a complex microbial community, it is difficult to explain the overall metabolic situation through the metabolism of individual bacteria. SCFAs, secondary bile acids (BAs), and tryptophan are three major microbial metabolites that take part in intestinal epithelial integrity and barrier function [93]. In particular, SCFAs, the main end metabolites produced in LBP fermentation, can regulate host physiology through multiple pathways: (i) lowering the local pH, lubricating the intestinal tract, promoting mucin secretion, and inhibiting the growth of pathogens and their adhesion to intestinal mucosa [94]; (ii) directly suppressing the activity of histone deacetylases (HDACs), which regulate the expression of inflammatory/immune genes, thus reducing the secretion of pro-inflammatory cytokines (e.g., TNF-a) [95]; (iii) activation of G protein-coupled receptors (GPCRs, such as GPR41, GPR43, and GPR109A) on the inner surface of epithelial cells or immune cells, thus triggering immune response in a very rapid manner [96]; (iv) acting as a major energy source for intestinal epithelial cells, promoting epithelial cell proliferation and differentiation, and improving intestinal epithelial barrier function [97]; (v) inhibition of the NF-κB signaling pathway and reduction of oxidative stress, thereby reducing colonic inflammation and even carcinogenesis [98,99]. Although SCFAs include acetate, propionate, n-butyrate, i-butyrate, n-valerate, and i-valerate, more than 90% of total SCFAs in the colon are constituted by the first three. Notably, numerous studies have shown that LBPs not only increase the concentrations of SCFAs, but promote the levels of SCFA-producing bacteria, such as acetate-generating Bifidobacterium, Prevotella, and Bacteroides [100,101,102], propionate-producing Bacteroides, Coprococcus, and Ruminococcus [62,78,90,92], and butyrate-producing Coprococcus and Faecalibacterium. [25,90,103]. In addition, our latest research showed that arabinogalactan-type LBP-3 could reverse the levels of certain specific amino acids (e.g., tryptophan, phenylalanine, lysine, glutamine, homoserine, and leucine) and organic acids, (e.g., kynurenine, 2-isopropylmalic acid, ascorbic acid, gluconic acid, (S)-2-hydroxyglutarate, and taurine) disturbed by DSS induction [104]. Moreover, pathway analysis indicated that the pentose phosphate pathway, phenylalanine, tyrosine and tryptophan biosynthesis, and phenylalanine metabolism were also altered by LBP-3 [104]. Additionally, LPS is also considered an intestinal bacterial metabolite, and its level was dramatically reduced by LBPs in HFD/streptozotocin (STZ)-induced diabetes in rats and mice [101,105]. Furthermore, urine metabolomics on an HFD/STZ-induced diabetic rat model revealed that administration of LBPs (glucan type) could enhance the levels of creatinine, 2,2,3-dihydroxybutyric acid, d-galacturonic acid, and citric acid, and reduce methylmalonic acid, benzoic acid, and xylitol, recovering them to normal levels [106]. In addition, supplementation with dietary Goji could decrease the contents of ω-6 polyunsaturated long-chain fatty acids (PUFAs, e.g., linoleic acid and arachidonic acid) and levels of the amino acids (l-valine, l-phenylalanine, l-serine, l-lysine, l-methionine, and l-glutamic acid) which were closely related with intestine inflammation in feces of interleukin (IL)-10-deficient mice [107]. Herein, the modulation of LBPs on gut microbiota and its metabolites in different experimental models (including fermentation of human gut microbiota in vitro) is summarized in Table 2. Considering the complex ecosystem of gut microbes, the alterations in microbial metabolites are probably not solely ascribed to LBPs. So knockout mouse models or isotope tracing methods need to be applied to understand the impacts of LBPs on microbial metabolism. More importantly, do the intermediate products (e.g., oligosaccharides) produced from the microbial degradation process of LBPs have benefits to the host? The oligosaccharide fragments liberated by polysaccharide-utilizing members (producers) are potentially available to other species unable to utilize polysaccharides alone (potential recipients) to form the ecological network of polysaccharide utilization among intestinal symbionts [119], which also makes it difficult to obtain active oligosaccharide fragments of LBPs. Furthermore, whether oligosaccharides produced by microbial degradation of polysaccharides have the ability to cross the vascular barrier into the systemic circulation, as well as their functional activities, are not yet known. It has been demonstrated that prebiotic GOS can improve mucosal barrier function by directly stimulating intestinal goblet cells [120,121]. In addition, a portion of oligosaccharides (e.g., GOS, human milk 2’-fucosyllactose, 6’-sialyllactose, and lacto-N-neotetraose) could be absorbed into plasma, thus reaching the systemic circulation [122,123], therefore, we speculated that oligosaccharides from LBP degradation by gut microbes may have access to the systemic circulation. At present, there are few reports about the active oligosaccharide fragments derived from microbial degradation of LBPs. Fortunately, our research group has obtained some oligosaccharides using a single culture of certain Bacteroides strains, and these oligosaccharides indeed contain GOS after derivatization with PMP and analysis by RP-HPLC-MS (unpublished data). Thus, more studies are needed to elucidate the degradation mechanism and explore the functions of these intermediate products (e.g., oligosaccharides). The effects of LBPs on gut microbiota and its metabolites, as well as intestinal barrier function, are shown in Figure 2. LBPs are predominantly fermented by intestinal microbes to produce favorable metabolites, especially SCFAs, and in turn, they also alter the microbial composition by promoting the proliferation of probiotics, inhibiting the growth of pathogens, and stabilizing commensal bacteria. As described above, the regulatory effect of LBPs on gut microbiota is diverse, owing to different physicochemical properties of LBPs, individual diversity in gut microbes, and even the conditions under different health states. The types of glycosidic linkage, monosaccharide composition, degree of polymerization, and branched chains of LBPs greatly determine the modulation of LBPs on the profiles of gut microbial communities. Hence, how the gut microbes utilize structurally specific LBPs needs to be further investigated using further in vitro and in vivo experiments.
Numerous studies of the gut microbial genome have so far broadened our understanding of the potential mechanisms underlying human diseases. The gut microbiota can impact host physiological functions and metabolism through promoting energy metabolism and regulating host/diet-derived compounds that alter host metabolic activity [1]. As mentioned in Section 3, the composition of microbial communities can be modulated by LBPs, and in turn, LBPs provide available substrates for fermentation by gut microbes. Dysbiosis of gut microbiota contributes to immune dysregulation, inflammatory responses, and various metabolic disorders in the host [1,66,124,125]. The diversity and composition of gut microbial communities play a crucial role on the maintenance of intestinal homeostasis, allowing the symbiotic fitness between gut microbiota and host immunity. Therefore, the interaction between LBPs and gut microbes is a potentially vital strategy to target host health benefits. The health effects of LBPs have been validated in both mice and human studies; however, the exact underlying mechanisms are still not fully understood. This review will summarize the intervention of LPBs on disease progression based on microbial strategies.
It has been demonstrated that central immune organs (e.g., bone marrow and thymus) and peripheral immune organs (e.g., spleen and intestinal lymph nodes) can be promoted by LBPs, thus enhancing host immunity [126]. However, an overreaction of the immune system can contribute to an uncontrolled inflammatory response and cytokine storm. Administration of LBPs could modulate the development and differentiation of immune cells such as T lymphocytes, B lymphocytes, macrophages, and dendritic cells (DCs), and downregulate the inflammatory immune response, inhibiting the secretion of pro-inflammatory cytokines [127]. Existing studies have primarily focused on the immune regulatory mechanisms of LBPs from a single immune enhancement or inflammatory inhibition instead of a bidirectional immune regulation. Herein, the bidirectional immune regulatory effects of LBPs were summarized and reviewed, of which the intestinal epithelial barrier function in host mucosal immune function has to be mentioned.
The intestinal epithelial barrier was mainly composed of a mucus layer, epithelial cells, and tight junctions (TJs) between epithelial cells [128]. The intestinal mucus layers, acting as the first line of defense against invading and symbiotic microbes, primarily consist of mucin-2 (MUC2) which is a glycoprotein with high-density clusters of O-linked glycans [129,130]. Reductions of Core 1 (Galβ1, 3GalNAcα1-Ser/Thr) and Core 3 (GlcNAcβ1, 3GalNAcαSer/Thr) O-glycans severely attenuate structural integrity and seriously disrupt intestinal mucosal barrier function, exacerbating microbial degradation of the mucus [131]. Furthermore, the most frequent aberrant glycosylations in inflammatory bowel disease (IBD) patients and animal models are the loss of Core 1 and Core 3 type O-glycans [131,132]. Insufficiency of nondigestible polysaccharides contributes to the erosion of the mucus layer by certain gut microbes (e.g., Akkermansia) which utilize mucin-type O-glycans as alternative nutrients, thus increasing invasion susceptibility of pathogens to intestinal epithelial cells and triggering immune and inflammatory responses [133,134]. Upon intestinal inflammation, a large number of inflammatory cytokines (e.g., IL-1β) and inflammatory mediators (e.g., iNOS) are secreted by intestinal mucosal immune cells (e.g., macrophages), which, in turn, damage intestinal epithelial cells, induce epithelial cell apoptosis, and reduce the expression of TJs, thereby compromising gut mucosal barrier function [135,136]. It has been demonstrated that MUC2-deficiency can lead to the development of spontaneous colitis with histologic damage, thinner mucus layer and increased permeability, which are susceptible to the invasion of epithelial cells by pathogens [137,138]. Our previous study found that the mucus layer got thicker and the expressions of mucin MUC2 and TJs (e.g., Claudin1 and ZO-1) were enhanced after supplementation with LBP-3 (arabinogalactan type) in DSS-induced colitis, thereby improving the intestinal barrier function [78]. In addition, administration of arabinogalactan-type LBPs could significantly elevate the levels of MUC2 and TJs (e.g., Claudin5 and Occludin1) and promote the number of goblet cells in both CTX-treated mice and normal mice [91,108]. Interestingly, one recent study found that sulfated polysaccharides from Gloiopeltis furcate could increase the abundances of potential probiotics Muribaculaceae and Roseburia, and enhance the levels of complex long-chain mucin O-glycans, especially sialylated G29 and G31 that contain eight to ten monosaccharides with two terminal N-acetylneuraminic acid residues, thus improving the intestinal barrier integrity and attenuating DSS-induced colonic mucosal damage [139]. Given that mucin O-glycans play a key role in host–microbiome interactions and that the glycan-peptide linkage of arabinogalactan-type LBPs was O-glycosidic linkage [22,39,105], arabinogalactan-type LBPs may protect colonic mucus layers by modulating the structure of gut microbiota, and in turn, intact mucin-type O-glycans enhance intestinal barrier function and prevent pathogen invasion. However, the protective mechanism of LBPs on mucin-type O-glycans needs further exploration.
The gut is the largest immune organ in the body, and it contains differentiated epithelial cells (e.g., enterocytes, goblet cells, and Paneth cells) and intestinal resident-immune cell subsets (e.g., B cells, T cells, DCs, and mesenteric-associated lymph nodes), which account for 70–80% of immune cells. Numerous studies have found that LBPs directly stimulate diverse immune cells or indirectly activate NF-κB signaling pathways through multiple pathways, thus promoting humoral and cellular immunity [10,17,126]. As mentioned in Section 3.1, only less than 1% of LBPs can pass through Caco2 monolayer cells [62], and it is perhaps these very limited LBPs that produce immune benefits, which is beyond the discussion in this review. In addition, Toll-like receptors (TLRs) of intestinal epithelial cells and immune cells (e.g., macrophages and DCs) can recognize pathogenic bacteria and their metabolites (e.g., LPS), which initiates signaling cascades including MyD88 and Interleukin-1 receptor-associated kinase (IRAK) and activates NF-κB signaling pathways, thereby releasing inflammatory mediators and activating the adaptive immune system [140]. Notably, intestinal immune responses are normally tolerant to commensals instead of pathogens in the steady state, and a healthy microbiome dynamically coexists with intestinal immunity in the co-evolution of host and microbe [141]. Consequently, the immune balance can be shaped by the composition of the microbial community. Many studies have found that the immune benefits of LBPs mediated by gut microbiota on the host are far beyond the gut, and impact the whole systemic immune responses. Arabinogalactan-type LBPS could improve thymus and spleen indexes and alleviate immune organ damage by enriching immune-related Lactobacillaceae, Bacteroidaceae, Verrucomicrobiaceae, and Prevotellaceae, as well as SCFAs in CTX-induced immunosuppressed mice [91]. Meanwhile, LBPs significantly upregulated the production of cytokines (e.g., IL-1β, IL-6, IL-2, IFN-γ, and TNF-α) by elevating the levels of Bacteroides and Gram-negative bacteria which contain LPS and activate the TLR4-MyD88-NF-κB signaling pathway [91]. Furthermore, arabinogalactan-pectin WBPPS improved immune function and regulated gut microbiota by increasing the abundances of Ruminococcaceae and Saccharimonadaceae in CTX-treated mice and lowering the levels of Tannerellaceae, Rikenellaceae, and Marinifilaceae, which were closely related to immune traits [92]. Similarly, another arabinogalactan-pectin LBP also exhibited immunoregulatory activity in CTX-induced mice by elevating splenic CD4+/CD8+ T-lymphocyte cell ratios and improving the diversity of gut microbiota, as well as the abundances of bacteria such as Bifidobacteriaceae, Rickenellaceae, and Prevotellaceae [62]. The above studies provide further insight into how LBPs with specific structures improve host immunity through gut microbiota: (i) the above mentioned LBPs exhibited immune enhancing properties, which were mainly composed of Ara, Gal and/or GalA with high branched chains; (ii) all arabinogalactan-type LBPs modulated gut microbiome structure by improving microbial diversity and upregulating the levels of probiotics (e.g., Lactobacillaceae) and commensal bacteria (e.g., Bacteroidaceae and Prevotellaceae), thus recovering them in a state of dynamic balance; (iii) in addition to the direct modulation of gut microbes, SCFAs, the prominent metabolites derived from microbial fermentation of LBPs, can be recognized by GPCRs on the surface of enterocytes or immune cells and involved in host immune response; (iv) LBPs promoted host immunity by directly improving central and peripheral immune organs (e.g., thymus and lymphatic) and indirectly enhancing the production of immune-related cytokines (e.g., IgA, TGF-β1, and TNF-α). Of note, the mechanistic evidence between immune and gut microbiota has been obtained mostly from animal models, and further research is needed to determine whether it can be applied to humans before relevant clinical trials [142].
An appropriate immune response protects the host from pathogenic infection, one overresponse can harm the host, and the inflammatory response is one outcome of an excessive immune reaction [143]. Dysbiosis in gut microbiota contributes to intestinal barrier dysfunction through impairing intestinal epithelial cells and enhancing permeability, and then endotoxins, pathogens, and other unfavorable molecules enter gut lamina propria, which can be recognized by TLR4 on macrophages or CD103+ dendritic cells, thereby triggering intestinal mucosal immune abnormalities [144]. Dietary supplementation with 1% LBP significantly ameliorated colonic mucosal damage, crypt destruction, and inflammatory infiltration, and increased the relative abundance of Lactobacillus and Butyricicoccus in DSS-induced colitis in wild C57BL/6 mice [145]. However, LBP failed to exert the protective effect against colitis, and fecal butyrate in the LBP group showed no difference compared to DSS treatment in germ-free mice [145]. These results indicated that LBP might alleviate colitis by modulating the composition of gut microbes, especially butyrate-producing bacteria, and gut microbiota seem to be essential for the anti-inflammatory activity of LBPs. In addition, acetate and propionate can inhibit HDAC and GPR43 signaling pathways, which contribute to the promotion of total colonic regulatory T cells (e.g., cTreg, Th1, and Th3) and production of anti-inflammatory cytokines IL-10 and transforming growth factor beta (TGFβ) [99]. Furthermore, our previous study also demonstrated that arabinogalactan-type LBP-3 exhibited an ameliorative effect against DSS-induced colitis by inhibiting the activation of TLR4-MyD88-NF-κB signaling pathways and reshaping the gut microbiota, as well as improving SCFA generation [78]. At present, most research focuses on the immune enhancing activity of LBPs, and less attention is paid to the immunosuppressive effects. Thus, future studies about anti-inflammation and its underlying mechanisms may be needed. In conclusion, both LBPs and their microbial metabolites, especially SCFAs, demonstrate bidirectional modulation of the immune response. LBPs modulate the host immune response by shaping gut microbiota and regulating the epithelial barrier function, thus establishing a symbiotic relationship of diet–host–microbiota (Figure 3). However, many aspects remain unclear in this symbiotic network: (i) How do LBPs regulate the gut microbes associated with the gut barrier and which bacteria taxa within the microbial community play a decisive role in the gut barrier? (ii) Apart from LBPs and the main metabolite SCFAs, do the intermediate product oligosaccharides have the bidirectional benefit of immunity? (iii) What is the molecular mechanism that causes LBPs to promote mucin secretion? Does it promote the proliferation of goblet cells or reduce the consumption of mucin by gut microbes? (iv) What is the effect of LBPs on the interaction between mucin O-glycosylation and gut microbiome? Future studies need to explore the above issues in depth and understand the protection mechanism of LBPs on the intestinal mucosal barrier.
Accumulating evidence has demonstrated that gut microbiota and its metabolites are crucial mediators in host energy metabolism, which participate in the progression of many metabolic diseases such as obesity, type 2 diabetes, and nonalcoholic fatty liver disease [5,146]. Although the etiology of metabolic syndrome (MetS) is still unclear, genetic inheritance, immunity, gut microbiota, and lifestyle may be responsible for the development of MetS [147]. Many studies have demonstrated that LBPs exhibited therapeutic effects on MetS, hence the role and mechanism of LBPs in the treatment of MetS were summarized and reviewed (Figure 4).
An expansion of Firmicutes and/or a drop in Bacteroidetes, i.e., an increased F/B ratio, which improves the capacity for the host to efficiently metabolize energy from nutrients, is usually observed in obesity and diabetes in both human and animal models [5,146,148]. Arabinogalactan-type LBP-W could significantly alleviate body weight and fat accumulation in HFD-induced obese mice and ameliorate the concomitant symptoms of hyperlipidemia and hyperglycemia, which are associated with the modulation of gut microbiota, such as improved diversity and richness, and reduced F/B and Proteobacteria (belonging to Gram-negative bacteria) [41]. It has been demonstrated that adipocytes can synthesize inflammatory cytokines such as TNF-a, IL-1β, and IL-6 and then accelerate inflammation in adipose tissue, which contributes to insulin resistance and other metabolic diseases such as type 2 diabetes [149]. Nevertheless, treatment with crude LBPs (without chemical characterization) recovered the gut microbiota dysbiosis by significantly elevating microbial diversity and beneficial bacteria (e.g., Bifidobacterium, Lactobacillus, and Alistipes) as well as their metabolites (e.g., SCFAs), and by reducing F/B ratio and opportunistic pathogens (e.g., Desulfovibrio, Deferribacteres, Tenericute, and Blautia) disturbed by STZ, consequently, effectively relieved the symptoms, such as fasting blood glucose (FBG) levels, serum triglycerides (TG), total cholesterol (TC), and plasma LPS levels in STZ-induced diabetes [101]. Interestingly, obesity may be closely related to certain specific bacteria such as Bifidobacterium, Lactobacillus, and Akkermansia, and these microbes are negatively correlated with obesity and type 2 diabetes [150]. Many studies have confirmed that LBPs can promote the proliferation of Bifidobacterium and Lactobacillus in vitro and in vivo [41,74,81]; however, few studies focus on the modulation of LBPs on Akkermansia in obesity [114]. Although these LBPs showed an amelioration effect on MetS, it is still challenging to further explore the potential molecular mechanisms, due to unclear key active components of crude polysaccharides and uncharacterized structures. Of note, a recent study showed that arabinogalactan-type LBPs significantly improved the levels of FBG, glycated hemoglobin, and pancreatic islet β-cell function in HFD/STZ-induced diabetic mice, and simultaneously discovered a key taxon (belonging to genus Allobaculum) associated with n-butyrate generation [105]. Furthermore, diabetic mice transplanted with LBPs-mediated gut microbiota had similar positive protection toward FBG (a decrease of 16.34%), however, such improvement could be deprived by antibiotics treatment [105]. The above studies suggested that LBPs could serve as a promising option for the treatment of type 2 diabetes based on the modulation of the intestinal microbial ecosystem.
The incidence of NAFLD varies from 20% to 30% in the general population and is as high as 75–100% in obesity [151,152]. Many studies have demonstrated that LBPs show protective effects on NAFLD by regulating gut microbiota. For example, intervention with arabinogalactan-pectin type WBPPS effectively improved CTX-induced hepatic tissue damage and oxidative stress by enhancing the activities of glutathione peroxidase (GSH-Px), superoxide dismutase (SOD), and catalase (CAT), and reducing the levels of malondialdehyde (MDA) and alanine aminotransferase (ALT) in the liver, which was closely associated with gut microbial composition, especially Ruminococcaceae, Saccharimonadaceae, and Tannerellaceae [92]. Similarly, administration of arabinogalactan-type LBP-W also could reduce HFD-induced hepatic steatosis, fat accumulation, liver inflammation, and cirrhosis [153]. In addition, the activation of hepatic TLR-4 by gut-derived LPS (via blood circulation) has been implicated in the pathogenesis of diet-induced NAFLD [154]. Meanwhile, glucan-type LBP also could reduce the activation of the LPS/TLR4/NF-κB signaling pathway via downregulating the harmful bacteria Enterococcaceae and its metabolites, LPS, in HFD-induced NAFLD rats, thereby reducing liver inflammation and lesions [79]. An important mechanism for the improvement of LBPs in diet-induced MetS may be that they promote the abundance of SCFA-producing microbiota (e.g., Lacticigenium, Butyricicoccus, and Lachnospiraceae_NK4A136_group) and simultaneously increase the levels of SCFAs, especially butyric and propionic acid [101,103]. Propionate and butyrate could prevent HFD-induced obesity by modulating free fatty acid receptors 2 and 3 (FFAR2 and FFAR 3) and gut microbes [155]. Furthermore, LBPs could significantly increase the level of n-butyrate and suppress the expression of pro-inflammatory cytokines by downregulating the expression of GPR43 and GPR109a and inhibiting the activation of the NF-κB pathway, thereby suppressing systemic obesity and chronic metabolic inflammation [79]. Subsequent studies also confirmed that LBP ameliorated obesity by modulating gut microbiota and SCFA production [41]. More importantly, butyrate and propionate are potent anti-obesity agents, particularly butyrate, playing a key role in the improvement of intestinal permeability and maintenance of gut microbial ecology [94]. In addition, the intestine and liver bidirectionally communicate through the gut–liver axis, which consists of the liver, gut and gut barrier [156]. As described above, LBPs also promote the expression of TJs to maintain gut barrier integrity [78,91,101] to ameliorate MetS. However, crude LBPs were currently employed to explore the protective effect on MetS in most studies, and the potential mechanisms still need further investigation, including: (i) which structural types of LBPs have the positive effect toward MetS, and the structure–activity relationship is unclear; (ii) investigation of the key bacteria and metabolites altered by LBPs is urgent, and the interaction between gut microbiota and its metabolites in MetS is unknown. We propose that future research should focus on the protective mechanism of LBPs with clear structures, provide a new therapeutic strategy for the prevention and treatment of MetS, and lay a foundation for in-depth study of the relationship between the structure and function of LBPs.
Apart from modulation of LBPs-mediated gut microbiota on host immune and MetS, such benefits have also been found in other diseases. For instance, supplementation with LBP (without chemical characterization) significantly improved lung inflammation and pulmonary edema through inhibiting the activation of the NF-κB pathways and cytochrome C in LPS-induced acute respiratory distress syndrome mice [157]. LBPs (without chemical characterization) could alleviate allergic asthma through reducing inflammatory cytokines (e.g., IFN-γ, TNF-α, IL-6, MCP-1, and IL-1β) in plasma and bronchoalveolar lavage fluid, and regulating gut microbiota, especially the improvement of beneficial Lactobacillus, Bifidobacterium, and Clostridiales [102]. In addition, LBE (without chemical characterization) significantly mitigated radiation-induced damage by increasing the potential beneficial bacteria Akkermansia and decreasing the relative abundance of harmful Rikenellaceae_RC9_gut_group, as well as modulating the corresponding metabolic pathways (e.g., tryptophan metabolism, indole alkaloids biosynthesis, D-arginine and D-ornithine metabolism, secondary bile acid biosynthesis, and arachidonic acid metabolism) [114]. Recent studies have demonstrated that the gut microbiota is involved in the regulation of emotions, behavior, and cognitive function through the gut–brain axis [158]. For example, LBP (without chemical characterization) may alleviate the emotional damage induced by chronic stress by improving alpha diversity, Lactobacillus, Prevotelaceae_ UCG-001, norank_f_Muribaculaceae, and SCFAs, thereby reducing the influence of stress factors on depressive damage in the offspring [113]. In addition, a clinical trial recently indicated that 300 mg d−1 LBP (without chemical characterization) could ameliorate depressive symptoms in adolescents with subthreshold depression, and demonstrated good tolerability with no adverse events [159].
The current review compiles the latest research findings on the isolation, purification, and structural types of LBPs, their modulation impact on gut microbiota, and the associated health benefits on host immunity and MetS. The composition of intestinal microbial communities is crucial for the utilization of LBPs which serve as the fermentation substrate and energy source for gut microbes to regulate gut microbial structure and metabolites. More importantly, the beneficial effects of LBPs on the host differ based on their diverse structural types and seem to be mediated by gut microbiota and its metabolites. In particular, SCFAs have been verified to modulate host immune responses and metabolic homeostasis. Although many studies have suggested that the health effects of LBPs are mediated by gut microbiota, in-depth studies are urgently needed to clarify the molecular mechanisms underlying immunity and MetS, and the following issues remain to be resolved: (i) The biological activities of LBPs have been investigated based on crude polysaccharides in most research, and it is difficult to reveal the molecular mechanism underlying the health effects due to their unclear structures. Meanwhile, another major limitation is a lack of standardization and quality control for the LBP used, which is adverse to subsequent clinical applications. (ii) What are the key gut microbes and enzymes in the degradation and utilization of LBPs? How do LBPs with specific structures shape the gut microbiota? The modulation of LBPs on intestinal microbiota is limited to simply analyzing the microbial diversity and abundance in current studies and the lack of microbial functions. (iii) LBPs could improve the intestinal epithelial barrier by mediating gut microbiota; however, what are molecular mechanisms by which LBPs increase mucin secretion? The interaction between the gut microbiome and mucin O-glycans is unclear. Final microbial metabolites, SCFAs, are involved in enhancing intestinal barrier function, regulating host immunity and metabolism, whether the intermediate products oligosaccharides have these benefits is unclear. More studies are needed to determine the metabolite profiles and their impacts on host health after supplementation with LBPs. (iv) LBPs are one of the most studied natural polysaccharides, which have great potential to provide safe and effective treatment for immune and metabolic diseases. However, the underlying mechanism between health effects and LBPs by mediating gut microbiota were mainly investigated in animal models, and large-scale clinical trials are needed to confirm the regulatory effects of LBPs in human immunity and metabolic diseases. In the future, exploring the biological functions of LBPs with diverse clear structures and the precise relationship between chemical structure–gut microbiota–biological activity of LBPs are urgently needed to provide a theoretical basis for how LBPs exert health effects on the human body, and lay a foundation for product development and clinical application. | true | true | true |
|
PMC9602844 | 36047802 | Shuqi Wu,Run Xu,Mengjiao Su,Can Gao,Yang Liu,Yujia Chen,Guangxin Luan,Xu Jia,Rui Wang | A pyrF-Based Efficient Genetic Manipulation Platform in Acinetobacter baumannii To Explore the Vital DNA Components of Adaptive Immunity for I-F CRISPR-Cas | 01-09-2022 | Acinetobacter baumannii,pyrF based,genetic manipulation system,CRISPR-Cas | ABSTRACT Acinetobacter baumannii is an important pathogenic bacterium with multidrug resistance which causes infections with high mortality rates. In-depth genetic analysis of A. baumannii virulence and drug-resistant genes is highly desirable. In this study, we utilized the conserved pyrF-flanking fragment to rapidly generate uracil auxotrophy hosts with pyrF deleted in model and clinical A. baumannii strains and then introduced the pyrF gene as the selectable and counterselectable marker to establish a series of gene manipulation vectors. For gene deletion with the suicide pyrF-based plasmid, the second-crossover colonies screened with the pyrF/5-fluoroorotic acid (5-FOA) system were obtained more quickly and efficiently than those screened with the sacB/sucrose system. By using the replicative plasmid, the recognized protospacer-adjacent motif (PAM) bias for type I-F CRISPR was experimentally revealed in A. baumannii AYE. Interestingly, interference recognized only the PAM-CC sequence, whereas adaptation priming tolerates 4 PAM sequences. Furthermore, we also performed a rapid and extensive modification of the I-F CRISPR-Cas elements and revealed that the role of double-nucleotide sequence mutants at the end of the repeat could be critical during both CRISPR interference and priming; we also found strong biases for A and demonstrated that adaptation could tolerate certain sequence and size variations of the leader in A. baumannii . In conclusion, this pyrF-based genetic manipulation system was readily applicable and efficient for exploring the genetic characteristics of A. baumannii . IMPORTANCE In this study, we developed the widely applicable and efficient pyrF-based selection and counterselection system in A. baumannii for gene manipulation. In most cases, this pyrF/5-FOA genetic manipulation system was very effective and enabled us to obtain marker-free mutants in a very short period of time. Utilizing this system and the separate mechanism of interference and/or primed adaptation, our experiments revealed some recognition mechanism differences for the key DNA elements of PAM, leader, and repeat in the priming adaptation process of the I-F CRISPR-Cas systems of A. baumannii , which provided some new and original insights for the study of the molecular mechanisms of these processes and laid a foundation for further studies. | A pyrF-Based Efficient Genetic Manipulation Platform in Acinetobacter baumannii To Explore the Vital DNA Components of Adaptive Immunity for I-F CRISPR-Cas
Acinetobacter baumannii is an important pathogenic bacterium with multidrug resistance which causes infections with high mortality rates. In-depth genetic analysis of A. baumannii virulence and drug-resistant genes is highly desirable. In this study, we utilized the conserved pyrF-flanking fragment to rapidly generate uracil auxotrophy hosts with pyrF deleted in model and clinical A. baumannii strains and then introduced the pyrF gene as the selectable and counterselectable marker to establish a series of gene manipulation vectors. For gene deletion with the suicide pyrF-based plasmid, the second-crossover colonies screened with the pyrF/5-fluoroorotic acid (5-FOA) system were obtained more quickly and efficiently than those screened with the sacB/sucrose system. By using the replicative plasmid, the recognized protospacer-adjacent motif (PAM) bias for type I-F CRISPR was experimentally revealed in A. baumannii AYE. Interestingly, interference recognized only the PAM-CC sequence, whereas adaptation priming tolerates 4 PAM sequences. Furthermore, we also performed a rapid and extensive modification of the I-F CRISPR-Cas elements and revealed that the role of double-nucleotide sequence mutants at the end of the repeat could be critical during both CRISPR interference and priming; we also found strong biases for A and demonstrated that adaptation could tolerate certain sequence and size variations of the leader in A. baumannii. In conclusion, this pyrF-based genetic manipulation system was readily applicable and efficient for exploring the genetic characteristics of A. baumannii. IMPORTANCE In this study, we developed the widely applicable and efficient pyrF-based selection and counterselection system in A. baumannii for gene manipulation. In most cases, this pyrF/5-FOA genetic manipulation system was very effective and enabled us to obtain marker-free mutants in a very short period of time. Utilizing this system and the separate mechanism of interference and/or primed adaptation, our experiments revealed some recognition mechanism differences for the key DNA elements of PAM, leader, and repeat in the priming adaptation process of the I-F CRISPR-Cas systems of A. baumannii, which provided some new and original insights for the study of the molecular mechanisms of these processes and laid a foundation for further studies.
Acinetobacter baumannii is a common and important pathogenic bacterium involved in hospital infections that causes wound infection, osteomyelitis, respiratory infections, bacteremia, etc. Since the 1980s, multidrug-resistant A. baumannii has become a challenge in the treatment of nosocomial infections; thus, the drug resistance mechanisms of A. baumannii have drawn constant attention. Over the last decade, the resistance rate of A. baumannii has been increasing, and strains resistant to carbapenems have been widely reported (1–4). Tigecycline and colistin were once considered the last antibiotics used to treat carbapenem-resistant A. baumannii (5, 6), but unfortunately, A. baumannii strains resistant to tigecycline and colistin have also appeared recently (7, 8). The virulence mechanisms of A. baumannii have also been a hot topic in recent years, as high mortality rates are observed in those infected in the hospital environment (9). Therefore, understanding the molecular mechanisms of drug resistance, cytotoxicity, and virulence of multidrug-resistant A. baumannii is crucial. Uncovering the mechanisms of A. baumannii drug resistance and virulence requires large amounts of genetic analysis. Previously, the lack of genetic manipulation tools and methods for studying these issues hindered the understanding of the diseases caused by A. baumannii. In the last several years, various putative methods for transforming exogenous DNA into A. baumannii have been developed and optimized, such as conjugation, electrotransformation, and natural transformation via motility (10–13). Recombinant cloning technology is a quick and efficient way to perform genetic manipulation in vivo which usually uses plasmids with antibiotic selection markers transformed into A. baumannii to create mutants (14). However, that brings antibiotic resistance genes into the genome, and repeated use of antibiotics for stress screening also causes the emergence of resistant colonies, so this method is not suited for multiple gene deletions. For multidrug-resistant A. baumannii strains, it was usually not easy to find a suitable antibiotic marker. To help address this problem, hygromycin B, which is an aminoglycoside antibiotic not used to treat infections in humans and has antimicrobial activity against a wide range of microorganisms, was applied to a diverse set of vectors (15). Later, Amin et al. demonstrated that a non-antibiotic resistance marker could be broadly used in most A. baumannii strains and developed the pMo130TelR vector, containing a tellurite resistance gene (16). In this study, we used the pMo130TelR vector as the selection marker to broadly generate A. baumannii hosts with pyrF deleted. This method does not use antibiotic selection for the gene deletion mutants and is beneficial for generating multiple gene deletions in a single strain. Recombination-mediated genetic engineering is one of the common methods to generate A. baumannii mutants. Usually, a second recombination event is initiated by the sacB gene as a counterselection cassette. The sacB gene expresses levansucrase in the presence of sucrose and is lethal in some Gram-negative bacteria, and this system was the most used in A. baumannii. However, when we used vectors with the sacB/sucrose system in A. baumannii, it always worked inefficiently and required a long time to subculture with sucrose for counterselection (15–17). This may be due to problems with the expression of the sacB gene or the function of levansucrase in A. baumannii. Tucker et al. (14) developed the recombinase-catalyzed homologous recombination method. This method can delete genes by a single-step transformation, but this brings antibiotic resistance genes into the genome, and the plasmids (RecAb-carrying and FLP-carrying plasmids) were both transformed and naturally removed twice, so there was a significant time cost (14). Researchers found that another type of non-antibiotic resistance selectable marker, i.e., auxotrophic markers such as amino acid auxotrophy, NADPH auxotrophy, thymidine auxotrophy, and uracil auxotrophy, is also a useful selectable tool in gene editing and vaccine production (17–21). Uracil auxotrophy has been successfully constructed in multiple strains, and some related genes are also widely used, such as pyrF, URA3, etc. (19, 22–24). The pyrF gene encodes orotidine-5′-monophosphate decarboxylase (25), which can convert orotidine-5′-monophosphate into UMP as a key intermediate for synthesis of uracil nucleotides (26), so it was widely used as a selectable marker in uracil auxotrophy pyrF mutants with medium lacking uracil. 5-Fluoroorotic acid (5-FOA) is an analogue of orotic acid which can be converted to a highly toxic compound (5-fluoro-UMP) by the product of pyrF and leads to bacterial death due to toxicity (24). Therefore, if the pyrF gene is deleted, the function of autonomously synthesizing uracil in the bacteria will be lost, and the mutants will be resistant to 5-FOA. Therefore, the pyrF gene was widely used as a selection and counterselection marker in uracil auxotrophy pyrF deletion mutants for gene manipulation (27–30). Moreover, to our knowledge, there are no previous reports of applying the pyrF gene as a selection and counterselection marker in A. baumannii. Clustered regularly interspaced short palindromic repeats (CRISPRs) and CRISPR-associated (Cas) proteins constitute an adaptive immune system against invading genetic elements in prokaryotes (31, 32). Different systems are currently divided into six types (I to VI), comprising several subtypes, which are mainly distinguished based on the existence of Cas genes and the proteins they encode. Thirty-seven percent of Acinetobacter genomes encode type I-F CRISPR-Cas systems, with some of the largest CRISPR arrays found so far in bacteria (33–35). The CRISPR-Cas immunity system in type I-F functions in three steps: CRISPR adaptation always needs a priming process to recognize the invader DNA and then takes up the target (protospacer) to integrate into the CRISPR array; in CRISPR expression, the CRISPR array is transcribed and processed by Cas6 proteins, and in CRISPR interference, the CRISPR RNAs (crRNAs) guide the Cascade (CRISPR-associated complex for antiviral defense) complex to recognize the protospacer and protospacer-adjacent motif (PAM) and then neutralize the invaders. The recognition is achieved through complementary interaction between the crRNA spacer and the protospacer and is also dependent on a specific short PAM (36, 37). CRISPR immunity makes it possible to effectively distinguish spacer sequences in the host CRISPR locus and identical protospacer sequences in the invading DNA, so as to avoid self-targeting (38–40). Therefore, a discrimination mechanism is required in the interference and adaptation process; otherwise, CRISPR DNA itself may become a potential target in the interference and adaptation process (41, 42). In type III-A systems, the “self versus nonself” discrimination mechanism identifies spacer DNA as a “self” component by recognizing the sequence matching between the repeat sequence present at the 5′ handle of crRNA and the repeat sequence in CRISPR locus, and other target DNA that does not match the sequence is identified as “nonself” (43). In the Haloarcula hispanica type I-B system, the primed adaptation, which attaches to and cooperates with the interference pathway, distinguishes target from nontarget by CRISPR RNA guidance and PAM recognition (39). These previous data demonstrated that the PAM and repeat end nucleotides preceding a spacer DNA were closely related to the discrimination mechanism (39, 43). Here, we systematically mutated PAM sequences and CRISPR repeat sequences at positions −1 and −2 and detected the effects of 16 PAM and repeat mutations for interference and priming of the adaptation process of the I-F system. The leader is also essential for adaptation, with different vital element functions. The leader sequence contains the promoter necessary to drive the transcription and encodes sequences recognized by the Cas1-2 complex and other cytokines, and importantly, it is required for insertion of repeat spacer units at the proximal leader edge of the repeat cassette (44). Previous studies that established the importance of leader sequences in the integration process point to the functional significance of the conserved DNA motifs (44–47). Therefore, it is necessary to explore the contributions and inherent specificity of CRISPR leader sequences in the adaptative immunity process. Here, based on the pMo130TelR plasmid, we used 5-FOA as the counterselectable agent and utilized the conserved pyrF flanking fragment to rapidly generate the uracil auxotrophy pyrF deletion hosts in model and clinical A. baumannii strains. Then, the pyrF gene of A. baumannii AYE was introduced as the selectable and counterselectable marker to establish a series of gene manipulation vectors. Utilizing pyrF-based suicide plasmids, we modified the I-F CRISPR-Cas system, and second-crossover colonies screened by pyrF/5-FOA system were obtained more quickly and efficiently than those screened with the sacB/sucrose system. Utilizing the pyrF-based replicative plasmid, we also detected the adaptation immunity of the I-F CRISPR-Cas system, highlighting the importance of the Cas gene and CRISPR DNA element, and the PAM sequences recognized in the priming and interference process. Based on the above-described tools, we constructed AYEΔF-PA with two separate CRISPR variants, namely, priming CRISPR (pCRISPR) and adaptation CRISPR (aCRISPR), to produce the priming crRNAs and to accept new spacers, respectively (47), and eventually generated a set of repeat end mutants and leader mutants for exploring the vital DNA elements for priming, spacer integration, and interference.
One purpose of this study was to develop a widely applicable pyrF-based selection and counterselection system in A. baumannii for high-efficiency genetic manipulation. As an initial step, we identified pyrF genes and their adjacent sequences with the Basic Local Alignment Search Tool (BLAST) algorithm. The only pyrF gene in each genome was predicted to encode orotidine 5′-phosphate decarboxylase and may be responsible for uracil biosynthesis and 5-FOA sensitivity. The pyrF genes and their adjacent genes are highly conserved, with identity at or above 98% in many A. baumannii strains, such as A. baumannii AYE, A. baumannii ATCC 19606, A. baumannii strain K09-14, A. baumannii strain 3207, and A. baumannii strain AC1633 (Fig. 1A), suggesting that these highly conserved up- and downstream flanking sequences may be used for homologous recombination to widely generate pyrF deletion hosts in A. baumannii. We chose to begin with the clinical model strain A. baumannii AYE because of its wide use, well-annotated genome sequence, and typical I-F CRISPR-Cas system. All strains and plasmids used in this study are listed in Table S1 in the supplemental material; 500-bp DNA sequences flanking pyrF were utilized for constructing pMo130TelR-ΔF (Fig. 1A and F). pMo130TelR-ΔF was integrated into the genome by first-crossover homologous recombination utilizing tellurite resistance as a selectable marker. The second-crossover colonies were screened with 5-FOA. After the first-generation screening, it was easy to create the ΔpyrF mutant by using the pyrF/5-FOA system, with an efficiency of 20/20 (Fig. 1B). PCR and DNA sequencing were then completed to confirm the pyrF gene deletion. With sucrose as a comparison, as described by Amin et al. (16), the plasmid was difficult to eliminate from the genome by the sacB/sucrose system, even after five generations (Fig. 1C). In LB medium, the growth of AYEΔF was significantly inhibited compared with that of A. baumannii AYE, whereas it grew as well as AYE with uracil added (Fig. 1D). In the uracil-auxotrophic medium M95, it could not grow, whereas it grew well when uracil was supplied. These results indicated that the deletion of pyrF gene limits the synthesis of uracil and subsequently affects growth, but growth could be complemented by the addition of uracil. The growth of the ΔpyrF strain was uninhibited in LB plates containing uracil and 5-FOA, but the wild-type strain could not grow (Fig. S1), which means that the ΔpyrF strain was resistant to 5-FOA. These results suggested that the pyrF gene was responsible for the uracil biosynthesis and 5-FOA sensitivity in A. baumannii AYE. To explore wider applications, we also easily deleted the pyrF gene in the A. baumannii model strains AYE and ATCC 19606 and 11 clinical strains by using the same method and same vector (pMo130TelR-ΔF) (Fig. 1B). As uracil-auxotrophic hosts, these ΔpyrF mutants also could not grow in synthetic M95 medium lacking uracil, whereas the growth of wild-type strains was uninhibited. When M95 plates had enough uracil added, the mutants also grew well (Fig. 1E). As mentioned above, the pyrF gene was expected to be broadly useful as a selection and counterselection marker for the genetic manipulation system in A. baumannii. The uracil auxotrophy and 5-FOA susceptibility were expected to be utilized for the selection (first crossover) and counterselection (second crossover), respectively.
For developing a pyrF-based positive selection and counterselection system, the pyrF-carrying plasmids were also needed as vector tools. The pyrF gene along with its 200-bp upstream promoter sequence was first introduced into pMo130TelR or pMo130 to generate pMo130TF or pMo130F (lacking the tellurite resistance marker), respectively, by replacing the sacB gene (Fig. 2A; Fig. S2A). pMo130TF and pMo130F are suicide plasmids for targeted gene deletion and insertion by homologous recombination. The A. baumannii ΔpyrF transformants with pMo130TF/pMo130F derivatives integrated into the genome can be positively selected by plating the cells on Tel/Km-containing medium or on synthetic medium without uracil, such as M95. On the other hand, cells in which the pMo130TF/pMo130F-derived sequence is eliminated from the genome can be counterselected by using medium containing 5-FOA. For gene expression, the autonomous replication plasmids pMo130TFR and pMo130FR were generated by introducing the rep gene of pWH1266 into pMo130TF and pMo130F, respectively (Fig. 2B; Fig. S2A). The transformation efficiency of pMo130TFR and pMo130FR increased sharply compared with that of pMo130TF and pMo130F in AYEΔF strains (Fig. 2A; Fig. S2B), suggesting that pMo130TFR and pMo130FR can replicate without integrating to the chromosome, while pMo130TF and pMo130F cannot replicate without the homologous sequence. The pMo130TFR and pMo130FR transformants all grew well on the M95 plates (Fig. 2B), which indicated that the pyrF gene in these plasmids could complete the uracil synthetic function and also can be used for positive selection in uracil-auxotrophic hosts. Therefore, these replicative plasmids can be autonomously replicated without integrating into the chromosome for gene expression, can be maintained by pyrF-based positive selection and may be eliminated by using medium containing 5-FOA as needed. Based on the autonomous replication plasmid pMo130TFR, green fluorescent protein (GFP) reporter plasmids and isopropyl-β-d-thiogalactopyranoside (IPTG)-induced expression plasmids were also derived. For the GFP reporter plasmid pMo130TFRG, the GFP-mut3 gene with a strong bacterial ribosome binding site (RBS), AAAGAGGAGAAA (48), was introduced, and Ptac as a promoter example (other researchers can change GFP-mut3 to different expression genes according to their experimental purpose when using this plasmid) was inserted into BamHI/PstI site before RBS-GFP to detect the promoter function. As shown in Fig. 2C, AYEΔF containing the plasmid pMo130TFRG-Ptac has a high ratio of fluorescence to optical density at 600 nm (OD600), whereas a very low ratio of fluorescence to OD600 was detected for the control without the promoter. This indicates that the GFP-mut3 gene was well expressed by the tac promoter and the GFP reporter plasmid works well. For the IPTG-induced expression plasmid pMo130TFRI, the LacI gene, Ptac, lac operator, and the RBS were introduced and GFP-mut3 as a gene example was inserted into BamHI/PstI site after the RBS. The IPTG-induced expression plasmid also worked well, and the GFP-mut3 was well repressed by LacI and also well induced by IPTG (Fig. 2D). Other researchers can change Ptac and GFP-mut3 genes to different promoter and expression genes according to their experimental purpose when using these plasmids. Therefore, these pyrF-based plasmids possessed a great capacity for gene manipulation as exceedingly convenient and versatile tools, with the pyrF gene as a selectable or counterselectable marker. The plasmids pMo130F and pMo130FR, with smaller sizes and greater loading capacity, may be used to insert or express large gene clusters. Selection of pyrF as the non-antibiotic resistance selection marker also could be a new choice.
The pyrF-carrying suicide plasmid and corresponding pyrF deletion-containing uracil-auxotrophic host constitute a new gene knockout system for A. baumannii. In order to demonstrate the convenience and usefulness of this system, we utilized it to modify the Cas gene and CRISPR sequence of the I-F CRISPR-Cas system in A. baumannii AYEΔF (Fig. S3B). pMo130TF-ΔCas1, pMo130TF-ΔCas3, pMo130TF-ΔCascade, and pMo130TF-aCRISPR were generated to delete Cas1, Cas3, Cascade, and a large part of the CRISPR except the leader and a single repeat structure (aCRISPR), as described above. These vectors were then transformed into AYEΔF, and the transformants integrated into the genome were positively selected by M95 or LB plates with Tel (30 mg/L). 5-FOA (50 mg/L) was then supplied as the counterselection for subsequent incubation, and 20 colonies were picked to streak onto replica plates for detection of the efficiency of excision of the plasmid-derived sequence from the genome by homologous recombination. For the pyrF/5-FOA system, the colonies with the plasmid-derived sequence removed were efficiency screened after the first generation, and as expected, almost all had the sequences removed, with efficiencies of 20/20, 20/20, 20/20, and 19/20 (Fig. 3B). PCR and DNA sequencing were also carried out to confirm the results, and the efficiencies of deleting Cas1, Cas3, Cascade, and a large part of CRISPR were 8/20, 7/20, 9/20, and 8/19 (Fig. 3C). As a comparison, we also utilized pMo130TelR (sacB/sucrose) to generate pMo130TelR-ΔCas1, pMo130TelR-ΔCas3, pMo130TelR-ΔCascade, and pMo130TelR-aCRISPR. However, it was difficult to isolate the mutant under selection with sucrose even after five generations (Fig. 1C and 3B). Therefore, the pyrF-based selection and counterselection system is an efficient and convenient choice for gene manipulation.
In the I-F CRISPR-Cas system in A. baumannii, the precise cutting of the target gene sequence by CRISPR-Cas depends on specific recognition of the PAM, and the spacer acquisition also needs to recognize the PAM for priming adaptation. To explore the differences in the PAM recognition mechanism during the interference and priming adaptation phases of the I-F CRISPR immune process, the replicative plasmid pMo130TFR, which has high transformation efficiency, was utilized to generate the target plasmid pMo130TFR-ENN-sp1. A schematic drawing of the strategy used for investigation is shown in Fig. 4A. Sixteen pMo130TFR-ENN-sp1 target plasmids for AYEΔF were constructed, all of which carried a sequence that is fully matched by spacer 1 (protospacer 1) and had a preceding double nucleotide as the PAM sequence inserted in the 5′ end of spacer 1. These plasmids were transformed into uracil-auxotrophic AYEΔF cells under selection pressure. As shown in Fig. 4B, transformation assays demonstrated that the PAM-CC produced strongly enhanced interference effects in strain AYEΔF. Although each plasmid carries the fully matched spacer 1, the interference effects of some PAM mutations were not obviously reduced. This suggests that the I-F CRISPR-Cas system functioned well, and the PAM-CC sequence appeared to be necessary for interference. However, when these transformants were subincubated and adaptation was detected by PCR, the priming process was detected for target plasmids with PAM-GC, -CC, -CT, and -TC (Fig. 4B). These results demonstrated that interference recognized PAM-CC almost exclusively, while priming adaptation could tolerance more PAM mutations and was able to recognize PAM-GC, -CC, -CT, and -TC. Based on the strongly enhanced interference effects of the replicative plasmid pMo130TFR-ECC-sp1, we utilized pMo130TFR-ECC-sp1 to verify the function of the above-described mutants with target gene deletions. pMo130TFR-ECC-sp1 and pMo130TFR (as the control) were transformed into A. baumannii AYEΔF, AYEΔFΔCas1, AYEΔFΔCas3, AYEΔFΔCascade, and AYEΔF-aCRISPR mutants. As shown in Fig. 4C, the target plasmid transformation efficiency for AYEΔFΔCas3, AYEΔFΔCascade, and AYEΔF-aCRISPR was similar to that for the control group, whereas the target plasmid transformation efficiency for AYEΔF and the Cas1 deletion mutant was obviously reduced. This suggests that almost all Cas proteins except Cas1 were involved in the interference process. The aCRISPR, with only the leader and repeat (without spacer 1), also could not detect the interference process, and therefore, priming is also necessary for adaptation in this I-F system.
In the native CRISPR system, a single CRISPR completes the processes of priming and spacer integration, which affects our research on the separate mechanisms of priming and spacer integration. Therefore, utilizing the pyrF/5-FOA genetic manipulation system, we established the AYEΔF-PA system with two CRISPR function systems for priming and spacer integration in AYEΔF (Fig. 5A). pCRISPR (containing a constitutive promoter and two repeat units flanking spacer 1 [pro-R1-sp1-R2]) was knocked into the genome to function as priming with plasmid pMo130TF-pCRISPR, while aCRISPR (containing the original leader and a single repeat) was modified to function as a new spacer integration site with plasmid pMo130TF-aCRISPR. Previous data reported by Li et al. (39) demonstrated that the repeat end nucleotides preceding a spacer DNA distinguish target from nontarget. For systematic analysis of the importance and base preference of the type I-F repeat end for interference and priming, we mutated the double-nucleotide end of the repeat 1 in pMo130TF-pCRISPR and then constructed 16 pCRISPR mutants (EP-NN) based on AYEΔF-PA. As shown in Fig. 5B, transformation assays with pMo130TFR-ECC-sp1 demonstrated that the interference effects were enhanced when the end of repeat 1 was AA, AT, AC, AG, TA, CA, or GA and that the AA nucleotide exhibited the most efficiently enhanced interference effects. The results revealed that the role of double-nucleotide sequence mutations at the end of repeat 1 could be critical during CRISPR interference and have strong biases for A. Then, the transformants were inoculated into M95 medium. We used PCR and sequencing technology to detect the occurrence of the integration process (Fig. 5B). The priming effects were also enhanced when the end of repeat1 contained A, while expansion was not detected or not obviously detected in the PCR analysis if the end of repeat 1 lacked A. The results revealed that the role of double-nucleotide sequence mutations at the end of repeat 1 also could be important for the priming function, and especially biases for A. These similar biases showed that there is a close link between the priming process and the interference process, and the Cascade protein participates in these two processes, in which the identification of repeat 1 termini is important. In the CRISPR immune system, the new spacer integration always occurred at the leader-proximal position during the adaptation process. Previous studies have shown that the leader motifs are very important for the identification of integration sites, but the nature of the recognition mechanism in type I-F CRISPR is not clear. Thus, we also truncated and mutated the leader sequence in pMo130TF-aCRISPR and then constructed aCRISPR mutants (EA-L) based on AYEΔF-PA, as shown in Fig. 5C. pMo130TFR-ECC-sp1 was also transformed into these mutants, and adaptations were detected by PCR and DNA sequencing. The EA-L2 mutation almost completely blocked the adaptation process. Obviously, L2 is rich in A bases, which showed that this sequence is critical for the adaptation process. We further constructed aCRISPR mutants containing various lengths of the leader sequence in aCRISPR, which were EA-L60, EA-L80, EA-L100, EA-L110, and EA-L120. The EA-L60 mutation completely and EA-L80 mutation almost completely blocked the adaptation process. EA-L60 and EA-L80 mutations both deleted the L2 region, which suggested that the L2 region of the repeat would be critical to the integration process in A. baumannii AYE. The sequence regions (L1 and L3 to L10) may affect the adaptation efficiency, but it will not completely block the integration process, indicating that this sequence is also important but not a key sequence.
Acinetobacter baumannii is an important pathogenic microorganism due to the high mortality rates of nosocomial infection and high levels of extremely drug-resistant isolates. Convenient and quick genetic methods can help investigate the basic biology of A. baumannii infection and transmission and analyze the molecular mechanisms of gene function for these processes. In this study, we developed a widely applicable and efficient pyrF-based selection and counterselection system in A. baumannii for gene manipulation. Strains containing pyrF deletions as the uracil-auxotrophic hosts and corresponding pyrF-carrying suicide vectors constituted the gene knockout system for A. baumannii. In most cases, this pyrF/5-FOA genetic manipulation system was very effective and enabled us to obtain marker-free mutants in a very short period of time, while the sacB/sucrose system for plasmid eliminating from the genome was still difficult even after five subgenerations. As a result, the selection (uracil auxotroph) and counterselection (5-FOA) markers can be utilized for developing gene manipulation systems for broad and convenient application in model and clinical A. baumannii strains. Moreover, a series of vectors for gene manipulation were established and applied to validate the convenience and versatility. A limitation of this method is that if this gene manipulation system is used in clinical strains, the efficiency may be lower than that in model strains for various reasons, such as low transformation efficiency or biofilm formation. Recently, Wang et al. (49) reported the development of a highly efficient and scarless genome engineering platform in A. baumannii by coupling a Cas9-mediated genome cleavage system with the A. baumannii RecAb recombination system. The CRISPR-Cas9-based genome editing system and the pyrF/5-FOA system can both be utilized for developing gene editing systems for use in diverse A. baumannii strains. Moreover, these genetic manipulation methods are time-saving and efficient, and in particular, they can yield marker-free mutants. In addition, the deletion efficiencies of the pCasAb-pSGAb system of Wang et al. are higher than those of the pyrF/5-FOA system. On the other hand, the pyrF/5-FOA system needs to transfer only one plasmid into the recipient cell and the plasmid is removed easily, which makes it more convenient than the pCasAb-pSGAb system. Based on the replicative pyrF-carrying plasmid, we explored the differences in the PAM recognition mechanism during the interference, priming, and acquisition phases of the I-F CRISPR immune process. Our results revealed that the PAM is also required to establish adaptive immunity in the 5′ end of identical protospacer sequences in the invading DNA and suggest that A. baumannii AYE interference almost exclusively recognizes the PAM-CC sequence, while priming adaptation could tolerance more PAM mutations and recognized PAM-GC, -CC, -CT, and -TC. Three of the four PAM sequences were generated by CC through a single point mutation, suggesting that these PAM mutations in the escape interference process drove the evolution of PAM selectivity in the adaptation process. The greater PAM selectivity during the adaptive immunity process explains the tolerance of PAM mutations in a target. Subsequently, based on the separate mechanisms of interference and/or primed adaptation, we mutated the double-nucleotide sequence at the end of repeat 1, which is transcribed as double nucleotides of the 5′ handle of crRNA. We observed the preference for A at the −1 and −2 positions in the interference and priming adaptation processes, which corresponds to native repeat motifs (the AA nucleotides occur at the end of repeat 1 of native I-F CRISPR-Cas in A. baumannii), which suggested that priming adaptation and interference can tolerate nucleotide variations within the crRNA molecule. These data exclude the possibility that the type I-F system makes use of a base pairing between the crRNA spacer and the protospacer to inhibit self-targeting. The priming adaptation and interference processes both require crRNA to target the protospacer based on base pairing and require the PAM verification mechanism to avoid targeting spacer DNA. These processes possess different tolerances for mutations in the PAM and the 5′ handle of crRNA at positions −1 and −2. We hope that providing the selection of PAM will establish a basis for furthering the understanding of the functional target and nontarget recognition mechanisms at the CRISPR site in A. baumannii. The Pseudomonas aeruginosa type I-F system targets foreign DNA through complex mechanisms, involving protein-mediated interaction with DNA and crRNA-guided interaction with cDNA and has been proposed as the model of the Csy complex target search process (50). Based on that and a model for a Cascade-mediated recognition mechanism proposed in a previous study (39), we envisioned a discrimination mechanism for CRISPR interference and priming adaptation. We speculated that the Cascade (CRISPR-associated complex for antiviral defense) complex may utilize its protein subunit(s) to distinguish target from nontarget by PAM scanning, and the interference and priming adaptation can avoid CRISPR DNA becoming a potential target, because the PAM is absent from repeat sequences in the host CRISPR locus and is present only next to complementary protospacer targets in foreign DNA. The binding affinities for Cascade complex with crRNA and the protospacer DNA with alternate PAMs results in the PAM and 5′ handle of CRISPR repeat having different nucleotide tolerances for the CRISPR interference and adaptation-priming processes, and we speculate that the double nucleotide at the end of repeat 1 (−1 and −2 positions in crRNA) may be responsible for regulating the proper anchoring between the 5′ handle repeat of crRNA and Cascade complex. After a PAM that allowed interference (PAM-CC) or adaptation (PAM-GC, -CC, -CT, and -TC) to occur was detected, the Cascade complex with crRNA was used to further examine whether the target sequence exactly matched the crRNA spacer sequence. When the protospacer exactly matches the crRNA spacer DNA, interference or adaptation will occur. The reasons for PAM selectivity and A affinity require a more sophisticated model and further experiments to explain. The leader motifs are also important for the identification of integration sites. To analyze the key recognition and acquisition mechanism in type I-F CRISPR, for gaining experimental insight into the comparative sequence analysis of I-F leader motifs in Acinetobacter, we mutated leader sequences and lengths that regulate CRISPR adaptation. Our results showed that the L2 region is rich in A bases and is critical for the adaptation process to acquire new spacers. This region was determined to be conserved by comparative analysis of the leader sequences of I-F CRISPR-Cas systems in Acinetobacter. Therefore, this observation reflected a specific conservative leader sequence pattern and regulation in the I-F CRISPR. We observed that a leader shortened to 60 bp does not support detectable acquisition. The high rate of spacer acquisition was obtained with at least 100 bp of the full leader sequence. The type I-E and I-F systems have been shown to rely on a DNA-bending protein called integration host factor (IHF). In type I CRISPR systems, identified leader DNA sequences are specifically recognized by the IHF protein to facilitate leader-proximal spacer integration (51). The mechanisms of CRISPR adaptation are diverse, but all systems that rely on IHF are expected to be phase dependent (51, 52). We hypothesized that the L2 region DNA in the leader may bind IHF protein sites, which creates a stabilized structure with the Cas1-2 integrase complex bound to the first repeat of the CRISPR locus and determines the appropriate integration site at the leader-repeat junction (44–46, 51, 53), but specific structural analysis is required to further investigate this. Moreover, the phase of leader motifs, rather than their distance from the leader-repeat junction, may be critical for efficient adaption in IHF-dependent systems. Our goal was to further clarify the mechanism of site-specific integration and provide a theoretical basis for limiting the transmission of virulence and drug resistance genes of Acinetobacter baumannii by the pyrF/5-FOA genetic manipulation system. This pyrF-based system was more suitable than antibiotics to modify the DNA motifs for exploring the vital DNA components of adaptive immunity in Acinetobacter baumannii I-F CRISPR-Cas, and it provides us with a feasible gene tool for gene manipulation in the laboratory. In conclusion, the pyrF-based system is an efficient and convenient system for broad use in most A. baumannii strains to study drug resistance, virulence, and other gene functions.
The strains, plasmids, and primers used in this study are listed in Table S1. E. coli strains were grown at 37°C in Luria-Bertani (LB) medium, supplemented with the appropriate agent. A. baumannii strains and mutants were grown at 37°C in LB medium, supplemented as needed with potassium tellurite (Tel; 30 mg/L), kanamycin (Km; 50 mg/L), uracil (U; 50 mg/L), 5-fluoroorotic acid (5-FOA; 50 mg/L), and agar (15 g/L). For pyrimidine-free medium, a synthetic medium (M95) was used, which comprised 0.6% Na2HPO4, 0.3% KH2PO4, 0.1%NH4Cl, 1%NaCl, 1 × 10−5% thiamine-HCl, 0.2% glucose, 1 mM MgSO4, and 0.5% Bacto Casamino Acids (pH 7.0).
The gene sequences analyzed in this study were obtained from NCBI (https://www.ncbi.nlm.nih.gov/); amino acid and nucleotide sequences in A. baumannii were analyzed through BLASTP and BLAST. All cloning steps were performed in E. coli DH5α. The primers used for vector construction and PCR detection are listed in Table S2. PCR was performed according to the manufacturer′s instructions by using Phanta Max master mix (Vazyme) for cloning and 2× Taq master mix (Vazyme) for detection. Restriction enzymes used for cloning were from New England BioLabs (Ipswich, MA). Plasmids were all constructed through homologous recombination technology using a ClonExpress II one-step cloning kit (Vazyme) or T4 DNA ligase (Vazyme).
(i) Construction of the plasmid for pyrF gene deletion. The pMo130TelR-ΔF vector was designed to delete the entire open reading frame of the pyrF gene in A. baumannii. The upstream and downstream 500-bp regions flanking pyrF in A. baumannii AYE were PCR amplified with the primers ΔpyrF-UF/UR and ΔpyrF-DF/DR and cloned into pMo130TelR (BamHI and PstI digested) to generate pMo130TelR-ΔF. (ii) Suicide plasmids with pyrF as a selectable and counterselectable marker. The pyrF gene (along with its 200-bp upstream promoter sequence) was PCR amplified with the primer pair pyrF-F/R. Subsequently, this fragment was inserted into the EcoRI/KpnI sites of pMo130TelR to replace the sacB gene for generating pMo130TF. Similarly, this pyrF gene was inserted into the EcoRI/KpnI sites of pMo130 to replace the sacB gene for generating pMo130F. The pMo130TF and pMo130F vectors were suicide plasmids for homologous recombination, which used the pyrF gene as the selectable and counterselectable marker. (iii) Autonomous replication plasmids with pyrF as a selectable marker. pMo130TFR and pMo130FR were autonomous replication plasmids, which did not need to integrate into the genome and could use the pyrF gene as a selection marker. The rep gene of pWH1266 was PCR amplified and cloned into pMo130TF (EcoRI digested) and pMo130F (EcoRI digested) to generate pMo130TFR and pMo130FR with the primer pair Rep-F/R. (iv) Reporter and induced plasmids with pyrF as a selectable marker. Based on pMo130TFR, we constructed the GFP reporter plasmid pMo130TFRG-Ptac and the IPTG-induced expression plasmid pMo130TFRI. To construct pMo130TFRG-Ptac, the tac promoter region and GFP-mut3 with a strong bacterial RBS gene, AAAGAGGAGAAA (54), were amplified with the primers Tac-GFP-F1/F2/F3/R and inserted into the BamHI/PstI sites of pMo130TFR. The lacI gene and lac operator were inserted into pMo130TFRG-Ptac (BamHI and SphI) to construct pMo130TFRI with the primers LacI-F/R. (v) Plasmids for gene manipulation. For gene deletion, the upstream and downstream 500-bp flanking regions of Cas1, Cas3, Cascade, and a large part of the CRISPR genes except the leader and a single repeat structure (aCRISPR) were PCR amplified and cloned into pMo130TF (BamHI and PstI digested) or pMo130TelR (BamHI and PstI digested) to generate pMo130TF-ΔCas1, pMo130TF-ΔCas3, pMo130TF-ΔCascade, pMo130TF-aCRISPR, pMo130TelR-ΔCas1, pMo130TelR-ΔCas3, pMo130TelR-ΔCascade, and pMo130TelR-aCRISPR with the primer pairs ΔCas1-UF/UR, ΔCas1-DF/DR, ΔCas3-UF/UR, ΔCas3-DF/DR, ΔCascade-UF/UR, ΔCascade-DF/DR, ΔCRISPR-UF/UR, and ΔCRISPR-DF/DR, respectively. (vi) Plasmids for CRISPR interference and adaptation detection. The pMo130TFR-ENN-sp1 plasmids are target plasmids for detecting CRISPR interference and adaptation. To construct pMo130TFR-ENN-sp1 target plasmids, nucleotide substitutions were performed by PCR mutagenesis. The PAM 5′-NN and spacer 1 of the AYE CRISPR-Cas system were PCR amplified with primer ENN-sp1-F/R and cloned into pMo130TFR (NotI and BamHI digested). The possible base composition of the PAM sequence was AA, AT, AC, AG, TA, TT, TC, TG, CA, CT, CC, CG, GA, GT, GC, or GG, yielding 16 different plasmids, each named in the format pMo130TFR-ENN-sp1, where NN represents the PAM sequence and 1 represents protospacer 1. E. coli S17-1 was transformed with the pMo130TFR-ENN-sp1 plasmids and pMo130TFR (as the control) and subsequently served as the donor strain in mating with AYEΔF.
Electrocompetent A. baumannii wild-type and mutant cells were prepared by inoculating a fresh colony and grown at 37°C overnight with 220 rpm agitation. One-milliliter cultures in the late logarithmic phase were centrifuged at 10,000 rpm and 4°C for 2 min. The harvested cells were washed three times with 10% cold glycerol and then resuspended in 100 μL of 10% cold glycerin for transformation. For electrical transformation, 500 ng of the autonomously replicating plasmids pMo130TFR and their derivative plasmids is needed for mixing with 100 μL of competent cells, while 10 μg of the suicide plasmids and their derivative plasmids is needed. The mixed cells were incubated on ice for 10 min and then transferred to a 1-mm cuvette at a voltage of 1,800 V, a capacitance of 25 μF, and a resistance of 200 Ω. After electroporation, the competent cells were transferred to a 1.5-mL microfuge tube with 500 μL LB liquid medium added and cultured at 200 rpm and 37°C for 45 min. The cells were plated on Tel-containing LB plates or M95 plates and cultured overnight at 37°C.
After electrical transformation, the cells were plated on Tel-containing LB plates or M95 plates to select the colonies that contained pyrF-carrying plasmids either integrated into the genome or autonomously replicated. For gene deletion or mutation by the pyrF/5-FOA system, excision of the plasmids integrated in the genome to screen the mutant colonies was performed by propagating in LB medium containing uracil (50 mg/L) and 5-FOA (50 mg/L) as a counterselection agent. For the sacB/sucrose system, we used sucrose (10%) as the counterselection agent to manipulate genes as previously described (16). The colonies were picked and streaked on replica plates (the same colony was streaked on LB plates and continuously streaked on LB plates supplemented with 30 mg/L Tel to confirm that the pyrF genes or target genes were eliminated from the genome) and then PCR amplified to screen the colonies with the correct size band and DNA sequencing.
The wild-type strains and mutants of A. baumannii were streaked on both LB plates and LB plates with 5-FOA (50 mg/L) and cultured overnight to test susceptibility to 5-FOA. For the uracil-prototrophic strains of A. baumannii, we measured the growth curve by liquid culture in both LB and LB with uracil (50 mg/L) and also streaking on M95 plates and M95 plates with uracil (50 mg/L).
The gene for the GFP-mut3 protein was linked to the tac promoter using the overlap extension PCR strategy. The pyrF-carrying plasmids were transformed into AYEΔF. For each transformation assay, three individual colonies were selected and cultured to the late exponential phase with or without 0.5 mM IPTG, and their OD600 and fluorescence were simultaneously determined using the Synergy H4 hybrid multimode microplate reader (BioTek, VT, USA). The fluorescence/OD600 ratio was calculated for each of the three individual samples, and averages and standard deviations were calculated.
We established a system with two CRISPR systems, priming CRISPR (pCRISPR) and adaptation CRISPR (aCRISPR) structures, the former for detecting the priming and interference process and the latter for containing the complete CRISPR leader to incorporate new spacers. An integrative plasmid, pMo130TF-pCRISPR, was first constructed, and it was transformed into AYEΔF-aCRISPR (AYEΔF with a large part of CRISPR deleted except the original leader and a single repeat structure) to yield AYEΔF-Pa with pCRISPR mutants and the original aCRISPR. The upstream homologous arm fragment (chromosomal sequences immediately upstream of Cas1) was amplified with the primers pCRISPR-FPF-SphI and pCRISPR-FPR-BamHI, and the downstream homologous arm fragment (chromosomal sequences immediately downstream of fragments of the ABAYE0994 gene) was amplified with primers pCRISPR-PPF-PstI/pCRISPR-PPR-NotI, while the promoter and two repeat units flanking spacer 1 (a variant pro-R1-sp1-R2 structure named priming-CRISPR) were amplified with the primers pCRISPR-F1-BamHI and pCRISPR-R1-PstI. The fragments were joined with suicide plasmid pMo130TF (NotI and SphI digested) to generate PCRI (Table S1) with T4 DNA ligase (Vazyme). Subsequently, the aCRISPR was designed to transform into AYEΔF-Pa to gain the AYEΔF-PA with a-CRISPR mutants. The chromosomal sequences upstream (aCRISPR-FPF-SphI/aCRISPR-FPR-BamHI) and downstream (aCRISPR-PPF-PstI/aCRISPR-PPR-NotI) of the AYE wild-type CRISPR, along with the leader and a single repeat (a variant leader R1 structure named adaptation CRISPR) with primers aCRISPR-F1-BamHI and aCRISPR-R1-PstI, were cloned into the suicide plasmid pMo130TF (NotI and SphI digested) to generate pMo130TF-aCRISPR by the same method. pMo130TF mutated double-nucleotide exchanges at the end of repeat 1 to yield the EP-NN mutants, which were transformed into cells for detecting the priming and interference process. pMo130TF-aCRISPR mutated the original leader motifs to yield the EA-L1~L10 and EA-L60~L120 mutants. All mutated CRISPRs were subjected to PCR mutagenesis and validated by PCR analysis and DNA sequencing with p130-F/R primers and were screened through the gene knockout strategy as described above.
The target plasmids were transformed into A. baumannii or its derivative strain by electrical transformation. For each target plasmid, at least three independent replicates were performed. For CRISPR interference detection, we counted the transformant colonies and conducted a comparative analysis.
To monitor spacer acquisition from the plasmid DNA, the transformants of each target plasmid were inoculated into M95 medium by passage culture, and the subgenerations were used as the colony PCR template (using the primer pair EXF/R or EACRI-F/R). PCR amplification of the sequence flanking the leader can detect the expansion. The PCR program consisted of the following steps: (i) 95°C for 3 min; (ii) 30 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 30 s; (iii) 72°C for 10 min. The PCR products were separated on a 2% agarose gel. The expanded bands (larger than the parental band) were produced by spacer insertion colonies (adaptation). | true | true | true |
PMC9602861 | Seong-Lan Yu,Yujin Kang,Da-Un Jeong,Dong Chul Lee,Hye Jin Jeon,Tae-Hyun Kim,Sung Ki Lee,Ae Ra Han,Jaeku Kang,Seok-Rae Park | The miR-182-5p/NDRG1 Axis Controls Endometrial Receptivity through the NF-κB/ZEB1/E-Cadherin Pathway | 14-10-2022 | endometrial receptivity,miR-182-5p,NDRG1,NF-κΒ/ZEB1/E-cadherin pathway | Endometrial receptivity is essential for successful pregnancy, and its impairment is a major cause of embryo-implantation failure. MicroRNAs (miRNAs) that regulate epigenetic modifications have been associated with endometrial receptivity. However, the molecular mechanisms whereby miRNAs regulate endometrial receptivity remain unclear. Therefore, we investigated whether miR-182 and its potential targets influence trophoblast cell attachment. miR-182 was expressed at lower levels in the secretory phase than in the proliferative phase of endometrium tissues from fertile donors. However, miR-182 expression was upregulated during the secretory phase in infertile women. Transfecting a synthetic miR-182-5p mimic decreased spheroid attachment of human JAr choriocarcinoma cells and E-cadherin expression (which is important for endometrial receptivity). miR-182-5p also downregulated N-Myc downstream regulated 1 (NDRG1), which was studied further. NDRG1 was upregulated in the secretory phase of the endometrium tissues and induced E-cadherin expression through the nuclear factor-κΒ (NF-κΒ)/zinc finger E-box binding homeobox 1 (ZEB1) signaling pathway. NDRG1-overexpressing or -depleted cells showed altered attachment rates of JAr spheroids. Collectively, our findings indicate that miR-182-5p-mediated NDRG1 downregulation impaired embryo implantation by upregulating the NF-κΒ/ZEB1/E-cadherin pathway. Hence, miR-182-5p is a potential biomarker for negative selection in endometrial receptivity and a therapeutic target for successful embryo implantation. | The miR-182-5p/NDRG1 Axis Controls Endometrial Receptivity through the NF-κB/ZEB1/E-Cadherin Pathway
Endometrial receptivity is essential for successful pregnancy, and its impairment is a major cause of embryo-implantation failure. MicroRNAs (miRNAs) that regulate epigenetic modifications have been associated with endometrial receptivity. However, the molecular mechanisms whereby miRNAs regulate endometrial receptivity remain unclear. Therefore, we investigated whether miR-182 and its potential targets influence trophoblast cell attachment. miR-182 was expressed at lower levels in the secretory phase than in the proliferative phase of endometrium tissues from fertile donors. However, miR-182 expression was upregulated during the secretory phase in infertile women. Transfecting a synthetic miR-182-5p mimic decreased spheroid attachment of human JAr choriocarcinoma cells and E-cadherin expression (which is important for endometrial receptivity). miR-182-5p also downregulated N-Myc downstream regulated 1 (NDRG1), which was studied further. NDRG1 was upregulated in the secretory phase of the endometrium tissues and induced E-cadherin expression through the nuclear factor-κΒ (NF-κΒ)/zinc finger E-box binding homeobox 1 (ZEB1) signaling pathway. NDRG1-overexpressing or -depleted cells showed altered attachment rates of JAr spheroids. Collectively, our findings indicate that miR-182-5p-mediated NDRG1 downregulation impaired embryo implantation by upregulating the NF-κΒ/ZEB1/E-cadherin pathway. Hence, miR-182-5p is a potential biomarker for negative selection in endometrial receptivity and a therapeutic target for successful embryo implantation.
The endometrium, the innermost lining layer of the uterus, is composed of a basal layer and functional layer, which provides an optimal environment for embryo implantation. In particular, cell–cell communication between the endometrial luminal epithelial cells of the functional layer and trophoblast cells of the blastocyst is important for successful implantation. Adhesion molecules, such as integrins, cadherins, and selectins, are differentially expressed in luminal epithelial cells during the endometrial menstrual cycle and play crucial roles in endometrial receptivity [1,2]. MicroRNAs (miRNAs) are small, single-stranded, non-coding RNAs that induce RNA silencing and the post-transcriptional regulation of gene expression via complementary base pairing [3,4,5]. Variable miRNAs are expressed at different endometrial stages during the menstrual cycle, as well as in several pathological gynecological conditions such as infertility, endometriosis, and preeclampsia [6,7]. Numerous reports have demonstrated a correlation between miRNA expression and endometrial receptivity [8]. miR-182 is a member of the miR-183 cluster, which includes miR-183, -182, and -96 [9]. miR-182 has been described as an oncogenic miRNA that targets multiple genes in pancreatic cancer, glioblastoma, lung cancer, breast cancer, and endometrial cancer. In some cancers, miR-182 promotes tumor metastasis by upregulating epithelial–mesenchymal transition (EMT)-related genes [10,11,12,13,14,15]. We previously reported that miR-182 was downregulated in the mid-secretory phase, between 20 and 24 days of the menstrual cycle, during the window of implantation (WOI) in the endometrium [16]. NDRG1 was expressed at higher levels in the secretory phase of the endometrium than in the proliferative phase [16,17]. Altmäe et al. (2017) suggested that NDRG1 acts as a transcriptomic biomarker for endometrial receptivity [18]. However, the relationship between miRNAs and NDRG1 in endometrial receptivity has not been clearly elucidated. Here, we focused on the role of the miR-182-5p/NDRG1 axis in regulating endometrial receptivity. We found that NDRG1 was downregulated by miR-182-5p and regulated endometrial receptivity by controlling the NF-κΒ/zinc finger E-box binding homeobox 1 (ZEB1)/E-cadherin signaling pathway in endometrial epithelial cells.
Previously, we performed next-generation sequencing to identify RNAs and non-coding RNAs related to embryo-implantation receptivity and found that miR-182-5p was more highly expressed in the proliferative phase than in the secretory phase during the endometrial receptivity period [16]. To validate the differential expression of miR-182-5p, we quantified miR-182-5p in the proliferative and secretory phases of normal endometrium tissues and during the secretory phase of infertility (Figure 1A). Consistent with our previous results, miR-182-5p was expressed at higher levels in the proliferative phase than in the secretory phase and was more highly expressed during infertility than in the normal secretory phase. Following this, to investigate the role of miR-182-5p in endometrial receptivity, we used human endometrial epithelial cell lines with different receptivities (receptive: RL95-2; non-receptive: AN3-CA). First, the JAr spheroid attachment served as an in vitro model for embryo implantation for both cell lines. The attachment percentage of JAr spheroids to RL95-2 cells was higher than that of AN3-CA cells, which verified that the RL95-2 cells were receptive endometrial epithelial cell lines, as suggested by Ho et al. [19] (Figure 1B). miR-182-5p was expressed at lower levels in RL95-2 cells than in AN3-CA cells (Figure 1C). Hence, we investigated the relationship between the degree of spheroid attachment and miR-182-5p expression. As shown in Figure 1D, spheroid attachment of JAr cells was attenuated in RL95-2 cells overexpressing miR-182-5p, suggesting that miR-182-5p expression correlated negatively with endometrial receptivity. To further investigate this possibility, we examined the expression pattern of E-cadherin, which plays an important role in cell adhesion. E-cadherin expression was downregulated by miR-182-5p overexpression (Figure 1E,F). These results suggest that miR-182-5p expression was inversely related to endometrial receptivity (which is required for successful embryo implantation).
Most miRNAs induce phenotypic changes by regulating the expression levels of target genes through post-transcriptional mechanisms [3]. To identify target genes associated with endometrial receptivity, we analyzed seven databases including Cupid, MirAncesTar, miRDB, MiRNATIP, MultiMiTar, RNA22, and TargetScan using the microRNA Data Integration Portal (mirDIP; http://ophid.utoronto.ca/mirDIP/; 5 February 2021). NDRG1 was selected as a target gene for miR-182-5p. Moreover, we previously reported a correlation between miR-182 and NDRG1 in a competing-endogenous RNA network [16]. Therefore, to investigate the association between endometrial receptivity and NDRG1, we evaluated NDRG1 expression in the proliferative and secretory phase of the normal endometrium and in the secretory phase of infertility. The mRNA-expression level of NDRG1 was significantly elevated in the secretory phase of the endometrium, compared to that in the proliferative phase, but did not differ significantly between the secretory phases of normal and infertile endometrium tissues (Figure 2A). However, NDRG1 protein abundance was lower in infertile versus normal endometrium tissues, although the difference was not significant (Figure 2B). In addition, NDRG1 mRNA and protein were upregulated in RL95-2 cells, which are receptive endometrial cell lines (Figure 2C,D). These data demonstrate the potential of NDRG1 as a putative target of miR-182-5p for regulating endometrial receptivity. Transfecting synthetic miR-182-5p mimics downregulated NDRG1 expression at both the mRNA and protein levels (Figure 2E,F). Moreover, NDRG1 mRNA expression was significantly increased by miR-182 knockdown (Figure 2G,H). Therefore, these results strongly suggest that miR-182-5p-mediated NDRG1 downregulation was associated with defective endometrial receptivity.
We performed in vitro implantation assays to investigate the correlation between NDRG1 expression and embryo implantation. The spheroid attachment of JAr cells was lower with NDRG1-depleted RL95-2 cells than with control cells (Figure 3A). In contrast, overexpressing NDRG1 in non-receptive AN3-CA cells significantly increased the JAr cell-attachment rate (Figure 3B). These results suggest that NDRG1 plays an important role in successful embryo implantation.
E-cadherin plays a critical role in endometrial receptivity, which is important for successful embryo implantation [20,21]. NDRG1 can upregulate E-cadherin expression in pancreatic cancer cells [22]. Therefore, to investigate the relationship between NDRG1 and E-cadherin, which were downregulated by miR-182-5p overexpression, we established an RL95-2 cell line with stably depleted NDRG1 expression using the short-hairpin RNA (shRNA)-mediated gene-silencing method. NDRG1 depletion markedly downregulated E-cadherin mRNA and protein expression (Figure 4A,B). We also established AN3-CA cells that overexpressed NDRG1. NDRG1 overexpression markedly upregulated E-cadherin mRNA and protein expression (Figure 4D,E). These data indicate that NDRG1 positively regulated E-cadherin expression in endometrial epithelial cells. To confirm that NDRG1 can regulate E-cadherin expression, we performed immunofluorescence staining with NDRG1-depleted or -overexpressing cells. E-cadherin expression was attenuated by NDRG1 depletion and enhanced by NDRG1 overexpression (Figure 4C,F). These results suggest that NDRG1 positively regulated E-cadherin expression, which is related to endometrial receptivity and is important for successful embryo implantation.
The NF-κΒ/ZEB1 pathway has been found to repress E-cadherin expression in cancer cell lines [23,24]. Therefore, to investigate the role of NDRG1 in the NF-κΒ/ZEB1 pathway, we analyzed the expression patterns of p65 and ZEB1 in NDRG1-depleted or -overexpressing cells. The mRNA-expression levels of p65 and ZEB1 were significantly elevated by NDRG1 depletion (Figure 5A,B). NDRG1 depletion also induced the protein expression of p65 and ZEB1, suggesting that NDRG1 may regulate endometrial receptivity through the NF-κΒ/ZEB1 pathway (Figure 5C). To test this hypothesis, we examined the expression levels of p65 and ZEB1 in AN3-CA cells overexpressing NDRG1. In contrast to NDRG1 depletion, both p65 and ZEB1 mRNA levels were downregulated by NDRG1 overexpression (Figure 5D,E). NDRG1 overexpression also reduced the protein-expression levels of p65 and ZEB1 (Figure 5F). In addition, the immunofluorescence-staining results agreed with the Western blotting results (Figure 5G–K). These findings suggest that NDRG1 may upregulate E-cadherin expression by downregulating the NF-κΒ/ZEB1 pathway in endometrial epithelial cells.
Endometrial receptivity is the first essential requirement for a successful pregnancy, and its disorder is one of the main causes of infertility due to implantation failure [25]. Several miRNAs have been associated with impaired endometrial receptivity. miR-30d, which is increased in the mid-secretory phase of the endometrium, may be associated with human endometrial receptivity [26,27,28]. Increased miR-30d expression can regulate target genes involved in proliferation and hormonal responses [29]. Shi et al. reported 105 differentially expressed miRNAs in the endometrium of patients with repeated implantation failure (RIF) to identify miRNAs related to impaired endometrial receptivity [30]. miR-200c was identified as a therapeutic target for infertility based on its association with impaired uterine receptivity [31]. It has also been reported that miR-543 downregulation is correlated with impaired endometrial receptivity during the WOI [32]. In addition, miR-183-5p and miR-149 were identified as regulators of endometrial receptivity reported previously [33,34]. Therefore, miRNAs are closely associated with endometrial receptivity. In this study, we generated evidence suggesting that miR-182 negatively regulates endometrial receptivity. miR-182 was expressed at lower levels in the mid-secretory phase of the endometrium than in the proliferative phase and was expressed at higher levels during infertility. Moreover, miR-182overexpression reduced JAr spheroid attachment in our in vitro implantation model (Figure 1A,D). E-cadherin expression was downregulated in RL95-2 cells overexpressing miR-182-5p (Figure 1E). Epithelial adhesion molecules play essential roles in successful embryo implantation. E-cadherin belongs to the cadherin superfamily and functions as a cell–cell adhesion molecule in the epithelium. Reardon et al. [35] demonstrated that cdh1 disruption results in the loss of uterine function and infertility, where embryos could not attach to the uterus in mice. In humans, E-cadherin expression is expressed at lower levels in the endometrial epithelium of infertile patients than in that of fertile women [36,37]. E-cadherin plays an essential role in embryo attachment to the endometrium, a process known as endometrial receptivity. Recently, it was reported that miR-182 overexpression upregulated EMT-related genes, such as N-cadherin, vimentin, and ZEB1, but downregulated E-cadherin expression in prostate cancer cells [38]. Consistent with previous studies, our data showed that miR-182 led to E-cadherin downregulation, indicating that miR-182 negatively regulates endometrial receptivity. Endometrial receptivity occurs during the mid-secretory phase of the menstrual cycle. In this study, miR-182 was downregulated in the mid-secretory phase of the endometrium, and its target gene, NDRG1, was upregulated when compared to expression observed during the proliferative and secretory phases of infertility (Figure 2A,B). Moreover, NDRG1 expression was expressed at higher levels in highly receptive RL95-2 endometrial cells than in non-receptive AN3-CA endometrial cells (Figure 2C,D). Moreover, NDRG1 overexpression improved JAr spheroid attachment in our in vitro implantation model (Figure 3). Meng et al. [39] demonstrated that NDRG1 expression was decreased in the uteri of aborted mice. These data indicate that NDRG1 may participate in the endometrial receptivity pathway. Previous findings have demonstrated that NDRG1 served a major role in inhibiting tumor metastasis via EMT inhibition. NDRG1 attenuated EMT and modulated E-cadherin expression by inhibiting caveolin-1 protein expression in colorectal cancer [40]. In addition, NDRG1 also upregulated E-cadherin expression by inhibiting the Smad2 pathway in nasopharyngeal cancer cells [41]. There were some limitations to the spheroid attachment experiment in this study. The spheroids derived from choriocarcinoma may not resemble with characteristics of embryo for implantation. In addition, endometrial epithelial and trophoblast cell lines may not indicate similarity with the cellular response in vivo. As mentioned above, NDRG1 positively regulated E-cadherin expression in endometrial epithelial cells (Figure 4). These results suggest that NDRG1 is closely related to E-cadherin-mediated endometrial receptivity, which is required for successful embryo implantation. In previous studies, NDRG1 inhibited the NF-κΒ signaling pathway, which attenuated E-cadherin expression in pancreatic and colorectal cancers [42,43]. E-cadherin expression was also downregulated by EMT-inducing transcription factors such as ZEB1 and ZEB2 [24]. As shown in Figure 5, we found that NDRG1 downregulated p65, which is related to the NF-κΒ pathway in endometrial epithelial cells. ZEB1 expression was also reduced by NDRG1 overexpression. These results indicate that NDRG1 enhanced endometrial receptivity by upregulating E-cadherin expression via inhibition of the NF-κΒ/ZEB1 pathway. In this study, we observed that miR-182-5p and its target NDRG1 regulated E-cadherin expression, which plays an important role in embryo implantation. The regulatory mechanism of E-cadherin expression was related to the NF-κΒ/ZEB1 pathway (Figure 6). miR-182-5p mimics led to impaired trophoblastic JAr spheroid attachment in our in vitro implantation model. Therefore, our data suggest that miR-182-5p may serve as a negative biomarker for embryo implantation and a therapeutic target for impaired endometrial receptivity.
Human endometrial tissues were collected from participants in the proliferative (9–11 menstrual cycle days; mcd) and secretory (20–24 mcd) phases at Konyang University Hospital, and infertility samples of the secretory phase (20–22 mcd) were collected from participants at MizMedi Hospital. Infertile patients did not receive medications, including hormone treatments, during the sample collection period. Endometrial sampling was performed using a disposable uterine sampler (Rampipella, RI.MOS, Mirandola, Italy). An experienced gynecological pathologist histologically determined the menstrual stages of the samples using Noyes criteria [44]. This study was approved by the Bioethics Committee of Konyang University Hospital (institutional review board [IRB] file No. KYUH 2018-11-007) and MizMedi Hospital (IRB file No. MMIRB 2018-3). The characteristics of the volunteers’ endometria are indicated in Table 1.
Human endometrial cancer cell lines (AN3-CA and RL95-2) were obtained from the American Type Culture Collection (Manassas, VA, USA). AN3-CA and RL95-2 were maintained in MEM and DMEM/F-12 (Hyclone, Logan, UT, USA) medium supplemented with 10% fetal bovine serum (FBS; Gibco, Waltham, MA, USA) and 1% penicillin-streptomycin (Hyclone, Logan, UT, USA). The cells were grown at 37 °C in a humidified atmosphere containing 5% CO2. The cells were transfected for 72 h with an miR-182-mimic or miR-182-inhibitor using RNAiMAX (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. AN3-CA cells were transfected with a pcDNA3.1-control or pcDNA3.1-NDRG1 expression vector using Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, USA), and RL95-2 cells were transduced with lentiviral particles expressing an shRNA against NDRG1 mRNA (Sigma Aldrich, St. Louis, MO, USA). NDRG1-overexpressing AN3-CA or NDRG1-knockdown RL95-2 cells were selected with neomycin or puromycin dihydrochloride after transduction, respectively.
Total RNA was isolated from cells and endometrial tissues using the TRIzol® reagent (Ambion, Austin, TX, USA; Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer’s instructions. To analyze Mrna expression, complementary DNAs (cDNAs) was synthesized from 2 μg (Cells) or 5 μg (tissues) total RNAs using Moloney Murine Leukemia Virus reverse transcriptase (Promega, Madison, WI, USA). qRT-PCR was performed with triplicate samples using iQ SYBR Green Supermix (Bio-Rad Laboratories, Hercules, CA, USA) and a CFX Connect Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA). The primers used for real-time PCR are shown in Table S1 of Supplemental Materials. We also identified single peaks in the melting curve of qPCR. The 2−ΔΔct method was used to calculate the relative mRNA-expression levels, using GAPDH as the internal control. To measure the relative miR-182-expression levels, cDNAs were synthesized with a reverse transcription miR-182-5p or RNU6B primer and the TaqMan miRNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer’s instructions. Quantitative miRNA expression was performed with TaqMan Master Mix II and TaqMan miRNA assay primers (Thermo Fisher Scientific, Waltham, MA, USA) using a CFX Connect Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer’s protocols. Relative miRNA-expression levels were calculated using the 2−ΔΔct method using RNU6B as an internal control.
The cells were lysed on ice using a radioimmunoprecipitation assay buffer (Jubiotech, Daejeon, Korea) containing protease and phosphatase inhibitors (Roche, Basel, Switzerland). Cell lysates were analyzed using a bicinchoninic acid assay (Thermo Fisher Scientific, Waltham, MA, USA). A 30 or 50 μg quantity of protein was separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and transferred to polyvinylidene difluoride membranes (Millipore, Burlington, MA, USA). The blots were incubated with 5% skim milk (Difco, Detroit, MI, USA) for 2 h at room temperature and then probed overnight at 4 °C with primary antibodies (Cell Signaling Technology, Danvers, MA, USA) against E-cadherin (1:1000 dilution, 3195), p65 (1:1000, 8242), ZEB1 (1:1000, 70512), NDRG1 (1:1000, 5196), or GAPDH (1:3000, 5174). On the following day, the blots were incubated with horseradish peroxidase-conjugated secondary antibodies (Millipore, 1:3000) and detected using an Enhanced Chemiluminescence Kit (Thermo Fisher Scientific, Waltham, MA, USA).
Immunofluorescence analysis was performed as previously described (Yu et al., 2019). Briefly, AN3-CA cells (transfected with pcDNA or a pcDNA-based NDRG1-overexpression vector) or RL95-2 cells (transduced with a control shRNA or an shRNA against NDRG1) were seeded onto coverslips in plates. On the following day, the cells were fixed with 4% paraformaldehyde (Sigma Aldrich, St. Louis, MO, USA) and permeabilized with 0.3% Triton X-100 (Sigma Aldrich, St. Louis, MO, USA). The cells were then blocked with 1% bovine serum albumin solution in phosphate-buffered saline and probed overnight with primary antibodies against p65, ZEB1, and E-cadherin (all from Cell Signaling Technology, Danvers, MA, USA). Following this, the cells were probed with secondary antibodies conjugated with Alexa Fluor 594 or Alexa Fluor 488 (Invitrogen, Waltham, MA, USA). Nuclei were stained with DAPI (Invitrogen, Waltham, MA, USA). A negative control was used to confirm the specificity of the antibodies used in this study (Supplementary Figure S1). Images were examined and captured using a confocal microscope (LSM710; Carl Zeiss, Oberkochen, Germany).
To prepare JAr cell spheroids, JAr cells were seeded into a V-bottom microplate (Greiner Bio-one, Kremsmünster, Austria) and incubated in DMEM containing 10% FBS (Gibco, Waltham, MA, USA) and 1% penicillin-streptomycin (Hyclone, Logan, UT, USA) for 24 h in a humidified atmosphere containing 5% CO2. Endometrial cells were cultured in separate wells of a 12-well plate until they reached 80% confluency. The spheroids were harvested and co-cultured on monolayers of AN3-CA cells (for 2 h) or RL95-2 cells (for 1 h) that were treated with the miR-182 mimic, NDRG1 shRNA, NDRG1-overexpression vector, or the respective control. To count the attached spheroids, the microplate was inverted and centrifuged. The attached spheroids were counted under a microscope (Olympus, Center Valley, PA, USA). The percentage of spheroid attachment was calculated as the proportion of attached spheroids after running the sample through an inverted centrifuge to the total number of spheroids. The implantation assay was repeated at least thrice.
All experiments were independently performed thrice, and the data are presented as the mean ± SEM. The results were analyzed using Student’s t-test or Mann–Whitney test. The thresholds for statistical significance were set at p < 0.05 and p < 0.01. | true | true | true |
|
PMC9603043 | Morgan Ramón-Landreau,Cristina Sánchez-Puelles,Noelia López-Sánchez,Anna Lozano-Ureña,Aina M. Llabrés-Mas,José M. Frade | E2F4DN Transgenic Mice: A Tool for the Evaluation of E2F4 as a Therapeutic Target in Neuropathology and Brain Aging | 11-10-2022 | acetylated E2F4,synapsis,tissue homeostasis,Alzheimer’s disease,5xFAD mice,neuroinflammation,microgliosis,reactive astrocytes | E2F4 was initially described as a transcription factor with a key function in the regulation of cell quiescence. Nevertheless, a number of recent studies have established that E2F4 can also play a relevant role in cell and tissue homeostasis, as well as tissue regeneration. For these non-canonical functions, E2F4 can also act in the cytoplasm, where it is able to interact with many homeostatic and synaptic regulators. Since E2F4 is expressed in the nervous system, it may fulfill a crucial role in brain function and homeostasis, being a promising multifactorial target for neurodegenerative diseases and brain aging. The regulation of E2F4 is complex, as it can be chemically modified through acetylation, from which we present evidence in the brain, as well as methylation, and phosphorylation. The phosphorylation of E2F4 within a conserved threonine motif induces cell cycle re-entry in neurons, while a dominant negative form of E2F4 (E2F4DN), in which the conserved threonines have been substituted by alanines, has been shown to act as a multifactorial therapeutic agent for Alzheimer’s disease (AD). We generated transgenic mice neuronally expressing E2F4DN. We have recently shown using this mouse strain that expression of E2F4DN in 5xFAD mice, a known murine model of AD, improved cognitive function, reduced neuronal tetraploidization, and induced a transcriptional program consistent with modulation of amyloid-β (Aβ) peptide proteostasis and brain homeostasis recovery. 5xFAD/E2F4DN mice also showed reduced microgliosis and astrogliosis in both the cerebral cortex and hippocampus at 3-6 months of age. Here, we analyzed the immune response in 1 year-old 5xFAD/E2F4DN mice, concluding that reduced microgliosis and astrogliosis is maintained at this late stage. In addition, the expression of E2F4DN also reduced age-associated microgliosis in wild-type mice, thus stressing its role as a brain homeostatic agent. We conclude that E2F4DN transgenic mice represent a promising tool for the evaluation of E2F4 as a therapeutic target in neuropathology and brain aging. | E2F4DN Transgenic Mice: A Tool for the Evaluation of E2F4 as a Therapeutic Target in Neuropathology and Brain Aging
E2F4 was initially described as a transcription factor with a key function in the regulation of cell quiescence. Nevertheless, a number of recent studies have established that E2F4 can also play a relevant role in cell and tissue homeostasis, as well as tissue regeneration. For these non-canonical functions, E2F4 can also act in the cytoplasm, where it is able to interact with many homeostatic and synaptic regulators. Since E2F4 is expressed in the nervous system, it may fulfill a crucial role in brain function and homeostasis, being a promising multifactorial target for neurodegenerative diseases and brain aging. The regulation of E2F4 is complex, as it can be chemically modified through acetylation, from which we present evidence in the brain, as well as methylation, and phosphorylation. The phosphorylation of E2F4 within a conserved threonine motif induces cell cycle re-entry in neurons, while a dominant negative form of E2F4 (E2F4DN), in which the conserved threonines have been substituted by alanines, has been shown to act as a multifactorial therapeutic agent for Alzheimer’s disease (AD). We generated transgenic mice neuronally expressing E2F4DN. We have recently shown using this mouse strain that expression of E2F4DN in 5xFAD mice, a known murine model of AD, improved cognitive function, reduced neuronal tetraploidization, and induced a transcriptional program consistent with modulation of amyloid-β (Aβ) peptide proteostasis and brain homeostasis recovery. 5xFAD/E2F4DN mice also showed reduced microgliosis and astrogliosis in both the cerebral cortex and hippocampus at 3-6 months of age. Here, we analyzed the immune response in 1 year-old 5xFAD/E2F4DN mice, concluding that reduced microgliosis and astrogliosis is maintained at this late stage. In addition, the expression of E2F4DN also reduced age-associated microgliosis in wild-type mice, thus stressing its role as a brain homeostatic agent. We conclude that E2F4DN transgenic mice represent a promising tool for the evaluation of E2F4 as a therapeutic target in neuropathology and brain aging.
E2 factor 4 (E2F4) is a member of the E2F family of transcription factors [1], which are primarily known to regulate the cell cycle. E2F4 was first described as a cell cycle repressor able to interact with p107 [2,3] and p130 [4], two members of the retinoblastoma (RB) family. Nevertheless, its capacity to repress cell cycle progression can be modulated, as it can also facilitate the cell cycle progression of cardiomyocytes, normal intestinal crypt cells, and colorectal cancer cells [5,6]. A number of reviews have been published describing the role of this transcription factor in quiescence and other cell cycle-related mechanisms [7,8,9], and we refer to the reader to these informative reviews for this aspect of E2F4 function. Interestingly, E2F4 can also play other important roles in cellular physiology, including cell and tissue homeostasis and tissue regeneration [7,8,10,11,12]. Therefore, E2F4 can be considered a multifactorial factor with an important impact on neuronal welfare and brain homeostasis [11,12], suggesting that it may be a promising candidate target for neurodegenerative diseases and brain aging. E2F4 is a phosphoprotein whose phosphorylation within an evolutionary-conserved threonine motif containing T248 (Figure 1) can modify its function [11,13]. This covalent modification has been targeted by substituting T248 and T250 with alanines, thus resulting in a dominant negative form of E2F4 (E2F4DN). This mutant form, or E2F4, prevents cell cycle re-entry in developing neurons [13] and is able to prevent Alzheimer’s disease (AD)-deleterious processes in 5xFAD mice [11], a murine model of this disease [14]. In this review, we will focus on the novel functions of E2F4 and their regulation as well as the covalent modifications of E2F4 that may modulate its function. We will also describe what has been published on E2F4DN transgenic mice, a mouse model generated in our laboratory that has been useful for the analysis of E2F4 as a multifactorial therapeutic factor for AD. Finally, we will describe how neuronal expression of E2F4DN reduces the neuroinflammatory response in both 5xFAD/E2F4DN double transgenic and wild-type (WT) mice at 1 year of age (i.e., middle-aged mice [15]).
Human E2F4 contains 413 amino acids (410 in mouse) distributed throughout four domains (Figure 1). As with other E2F members, it forms a heterodimer with dimerization partner (DP) proteins through its dimerization domain (DD), located at the N-terminus of the molecule. The DD domain is required for its interaction with DNA through the DNA-binding domain (DBD). A third domain located at the C-terminus is required for the function of E2F4 as a transcription factor [16]; this transactivation domain (TD) is blocked when the retinoblastoma (RB) family proteins p107 or p130 interact with E2F4 through its protein-binding domain [10]. This interaction is crucial for the control of the E2F4 function as a transcription factor. Finally, E2F4 has a region that has been proposed as a regulatory domain (RD) [13] in which phosphorylatable residues, such as T248 (see below), are placed. In addition, two nuclear export signals (NES) are present in E2F4, one located within the DBD and the other in the DD [17,18]. These sequences maintain E2F4 within the cytoplasm unless it interacts with p107 or p130, which are required for the translocation of E2F4 to the nucleus. In addition, other factors can induce the translocation of E2F4 to the nucleus, as the latter can also regulate transcription through RB-independent mechanisms [19]. In this regard, E2F4 can interact with KPNB1, RanGAP1, and RanBP2 [19], three proteins that are involved in nuclear import [20,21], and may facilitate E2F4 nuclear translocation in the absence of RB family members. Moreover, E2F4 may be translocated to the nucleus with the help of DP family members DP-2 and DP-3 [22,23,24], likely due to the presence of a nuclear localization signal (NLS) in their sequence, as has been shown in DP-2 [25]. Finally, E2F4 harbors a weak putative NLS in amino acids 52-61 [25], suggesting that E2F4 can translocate into the nucleus in a cofactor-independent manner, similar to E2F5 during keratinocyte differentiation [26]. A ChIP-seq analysis performed in human lymphoblastoid cells identified around 16,000 E2F4 binding sites that potentially regulate 7346 target genes with wide-ranging functions, including cell cycle regulation, DNA repair, RNA processing, stress response, apoptosis, ubiquitination, protein transport and targeting, protein folding, and I-κB kinase/NF-κB cascade regulation [27]. In these cells, E2F4 can bind near transcription start sites (TSSs), a finding confirmed by others [28]. In addition, functional distal sites for E2F4 can be located more than 20 kb away from the annotated TSSs. In both cases, E2F4 can act as an activator as well as a repressor [27]. This analysis also indicated that E2F4 can bind to the promoters of 780 transcription factors, suggesting that E2F4 can indirectly regulate broad classes of genes [27]. Other authors have confirmed that E2F4 can bind to genes related to DNA repair, DNA damage, and G2/M checkpoints, as well as to other classical functions, such as cell cycle regulation, DNA replication, chromosome transactions, and mitotic regulation [29]. In most cases, E2F4 can bind to a specific promoter together with other members of the E2F family [28], indicating that the E2F4 function is subjected to complex cross-regulatory networks [30,31]. Many E2F4 binding sites have been analyzed in specific gene regulatory regions [32]. For instance, the release of a p130-E2F4 complex from sequences immediately upstream of the transcription initiation site of the human CDC2 promoter has been shown to coincide with the induction of CDC2 expression [33]. Several lines of evidence indicate that E2F4 is able to control complex transcriptional regulatory networks in specific cells, thus supporting its multifactorial capacity as a transcription factor. For instance, a combined analysis using gene ontology and expression data has been used to define the network controlled by E2F4 in B cells [34]. In addition, loss-of-function studies on E2F4 silencing using a specific shRNA in acute myeloid leukemia cells have revealed that 276 genes show altered expression patterns in these cells [35]. These E2F4-regulated genes are mostly involved in the regulation of the mitogen-activated protein kinase (MAPK) signaling pathway. The regulation of gene transcription by E2F4 seems to be mediated through histone acetylation, as E2F4 may interact with CREB binding protein (histone acetyltransferase) [19], and sites where E2F4 binds are histone-modified [27].
E2F4 lacks a strong NLS, which suggests that this protein could play a significant role in the cytoplasm [36]. This is, for instance, the case of the regulation of centriole amplification during multiciliogenesis, which is mediated by the interaction of E2F4 with Deup1 and SAS6, two components of the centriole replication machinery [37]. Indeed, cytoplasmic E2F4 forms organizing centers in multiciliated cells [38]. While centrioles are known to undergo one round of duplication per cell cycle in normal proliferating cells, multiciliated cells show a massive assembly of these organelles during G0, a process initiated by Multicilin in combination with E2F4 (or E2F5) and Dp1 [39,40,41]. The capacity of E2F4 to function out of the nucleus is consistent with a study by Hsu and collaborators [19]. These authors identified a number of E2F4 interactors in mouse embryonic stem cells (mESCs) and a retinal pigment epithelium (RPE)-derived cell line of human origin [19]. Several of these interactors are located outside of the cell nucleus since a cellular component (CC) ontology analysis performed by us using the E2F4 interactors described by Hsu and collaborators [19] confirmed that E2F4 may be functional in the cytoplasm of mESCs (Table S1) and both cytoplasm and extracellular vesicles from RPE-derived cells (Table S2).
Proteins can be posttranslationally modified through covalent processing events that change their properties, either by proteolytic cleavage or by adding a modifying group, such as acetyl, phosphoryl, glycosyl, and methyl, to one or more amino acids [42]. More than 400 different types of posttranslational modifications [43] affect many aspects of protein function. Some of these chemical modifications have been described in E2F4. As in the case of other regulators of the cell cycle, E2F4 can be ubiquitinated as a mechanism regulating its protein levels [44]. In addition, E2F4 activity could be modulated by protein acetylation, as observed with another member of the E2F family of transcription factors, E2F1 [45]. E2F1 can be acetylated in sites that lie adjacent to the DBD, thus increasing its DNA-binding ability and activation potential, as well as its protein half-life [45]. In the case of E2F4, Hsu and collaborators [19] demonstrated that both human and mouse E2F4 can be significantly acetylated in K37 and K96. These residues are located within the DBD and DD, respectively, thus suggesting that the capacity for DNA binding and DP heterodimerization of E2F4 can be compromised. This may facilitate the cytoplasmic function of E2F4 as a multifactorial protein. These authors also found small levels of acetylation in K20, K28, K44, K73, K82, K101, K177, K197, K230, and K347 from human E2F4 and in K28, K44, K101, K118, K177, K178, and K339 from mouse E2F4. Most of these residues are located within the DBD and DD of E2F4, suggesting that their acetylation can also participate in the regulation of DNA binding and the DP heterodimerization of E2F4. Using an acetylated K96-specific antibody, we verified that K96 becomes acetylated in some structures of the adult mouse brain in vivo (Figure 2). This form of acetylated E2F4 can be detected in NeuN-positive cells (i.e., neurons) within the hippocampus (dentate gyrus) (Figure 2a), cerebellum (Figure 2b), and NeuN-negative cells located in the rostral migratory stream (RMS) (Figure 2c), likely neural progenitors. Some NeuN-negative cells in the cerebellum also showed acetylated E2F4-specific immunoreactivity (Figure 2b). In non-histone proteins, methylation represents a chemical modification participating in diverse processes, such as cell cycle control, DNA repair, senescence, differentiation, apoptosis, and tumorigenesis [46]. As a multifactorial factor, E2F4 can also become methylated. In this regard, Hsu and collaborators [19] have shown that a significant proportion of K73, K197, and R357 (R360 in mice) residues from E2F4 can be methylated. Interestingly, the methylation of K197 in E2F4 is reminiscent of a similar process in E2F1, affecting K185, which is involved in the regulation of E2F1-induced cell death [46,47,48]. Other residues of human (K20, K37, K53, K57, K74, K96, K101, R147, K177, K230, and K347) and mouse (R297 and K339) E2F4 can also be methylated, as reported by Hsu and collaborators [19]. Finally, the most prominent mechanism regulating E2F4 activity is protein phosphorylation. E2F4 has several residues susceptible to phosphorylation (Figure 1), and several lines of evidence indicate that E2F4 can undergo phosphorylation [49] to modulate its function. In this regard, this chemical modification may regulate E2F4-mediated transcription, either by disrupting its DNA-binding ability, as observed in 3T3 cells [50], or by enhancing the DNA binding of the E2F4/p130 repressor complex, as demonstrated in human fibroblasts [51]. Seven of the theoretical phosphorylation sites of E2F4, including T14, S202, S218, T224, S244, T248, and S384, have been demonstrated to become phosphorylated [52]. Other authors have confirmed the phosphorylation of T14, S218, S244, T248, and S381 in human E2F4 [19], of S218, T224, T249, and S384 in mouse E2F4 [19], and the ortologue of T248/T250 (T261/T263) in chicken E2F4 [13]. In addition, phosphorylation of E2F4 in T249 has been observed in mouse brain extracts using a phosphosite-specific antibody [11], and indirect evidence for the phosphorylation of T248 in the human brain was obtained using a proximity ligation assay with anti-E2F4 and anti-phosphothreonine antibodies [12]. Hsu and collaborators [19] also found evidence of phosphorylation in S16, Y139, S185, S187, S220, S223, and Y389 from human E2F4 and in S75, Y139, T153, S223, S240, T266, Y392, and Y394 from mouse E2F4. We will further discuss the effects of E2F4 phosphorylation in Section 5.2.
In addition to its classical function in regulating quiescence in proliferating cells, E2F4 can also participate in several homeostatic processes. For instance, E2F4 has been associated with the DNA damage checkpoint and repair pathways [29,53,54] (see below), prevention of DNA damage-associated cell death [31], repression of apoptotic genes [55], epigenetics [56], metabolism regulation [57,58], autophagy [59], inflammation [60], and cell repair [61]. In addition, E2F4 function has been associated with oxidative stress [62]. In this regard, the p107-E2F4 complex downregulates PGC-1alpha expression [63], an enzyme that protects cells against oxidative stress and reduces mitochondrial dysfunction in AD [64,65]. The ability of E2F4 to regulate several homeostatic functions may have evolved from its capacity to regulate processes primarily associated with cell cycle arrest and cell differentiation. Indeed, under growth arrest conditions, E2F4 can repress a common set of genes involved in mitochondrial biogenesis and metabolism [66]. Moreover, E2F4 participates in the differentiation of multiple cell types, including the differentiation of myocytes [22,36,67,68,69], neural cells [30,70], adipocytes [71,72,73,74], hematopoietic cells [75], chondrocytes [76], extra-embryonic tissues [77], endothelial cells [78], epithelial cells [79], and multiciliated cells [80,81]. E2F4 can also regulate eye and brain patterning [82,83,84,85], as well as endocytosis and water channel transport in the testes [81]. The capacity of E2F4 to act as a multifactorial factor is likely mediated by the different interactors to which this molecule can bind. In this regard, E2F4 can perform non-canonical actions in cells in the absence of RB family proteins, allowing the transactivation domain to interact with other proteins [19]. After performing biological process (BP) ontology analysis, we found that many E2F4 interactors identified by these authors are related to non-cell cycle processes, including DNA repair, stem cell population maintenance, protein sumoylation in mESCs (Table S3), as well as retina homeostasis, RNA splicing, organ regeneration, and regulation of lipid kinase activity in RPE-derived cells (Table S4).
Cells have to constantly respond to genotoxic insults that may induce DNA modifications, which usually lead to genome instability. Accumulation of damaged DNA is deleterious for cells since it often results in abnormal proliferation, cell aging, or cell death. Eukaryotic cells have acquired mechanisms of defense against this damage; globally, they are referred to as DNA damage response (DDR), which are in charge of monitoring and removing lesions in their DNA [86]. In this regard, mammalian cells are equipped with several DNA repair pathways, which can be classified into two main groups [87]. On the one hand, the machinery involved in base excision repair, nucleotide excision repair (NER), and mismatch repair can fix single-strand mutations. On the other hand, double strand breaks (DSBs) can be repared through two main mechanisms: homologous recombination (HR), which repairs DSBs during the S-phase or G2 since the sister chromatic is used as a template, and non-homologous end-joining (NHEJ), which is able to repair DSBs at any stage of the cell cycle and in quiescent and postmitotic cells. DDR can be transcriptionally regulated by E2F factors. These transcription factors usually bind to two adjacent E2F sites within the regulatory regions of genes involved in DNA damage checkpoint and repair [88], thus allowing for functional interactions. Two known E2F factors regulating DDR are E2F4 and E2F1 [27,29], which functionally counteract each other. For instance, E2F4 silencing in MCF7 epithelial breast cells treated with benzoapyrene, an environmental pollutant that triggers DNA damage [89], results in E2F1 derepression and the subsequent induction of DNA repair factors [90]. In primary neurons, the repair response to DSBs is also regulated by E2F1 and E2F4. In this cellular system, E2F1 enhances Cited2 expression, a pro-apoptotic gene required for delayed neuronal cell death, while E2F4 strongly inhibits Cited2 transcription, helping to cell survival [31]. Finally, E2F4 has also been implicated in NER since the p130/E2F4 complex controls the expression of xeroderma pigmentosum complementation group C [53], a protein that serves as the primary initiating factor in the global genome NER pathway [91]. There is also evidence that hypoxia and the anti-angiogenic agent cediranib are both able to induce the binding of p130/E2F4 complexes to E2F consensus sequences in the promoters of homology-directed DNA repair genes, thus reducing gene expression [54,88,92]. In most paradigms, E2F4 seems to act as a repressor of genes involved in DNA damage checkpoint and repair. This function may be favored by the stress kinase p38MAPK, which phosphorylates E2F4 [13] and becomes activated by the DDR [93]. Therefore, the expression of a non-phosphorylatable form of E2F4 (E2F4DN) might modulate the maintenance of the expression of genes involved in DDR.
E2F4 has been related to cognitive impairment [94] and the pathogenesis of AD [95], as well as to other neurological diseases [96], including Parkinson´s disease/mild cognitive impairment [97]. Since AD is largely a synaptic failure [98] occurring prior to cognitive decline or cell death [99], it can be speculated that E2F4 is important for synaptic function.
E2F4 has the potential to regulate the expression of an ample number of synaptic proteins. As evidenced by ChIP-seq datasets from the ENCODE transcription factor targets dataset interrogated with the Harmonizome tool [100], E2F4 can bind to 46 synaptic protein-encoding genes (Table S5), as well as 127 genes encoding for ion channel subunits (Table S6). In this regard, there is direct evidence that E2F4 can regulate synaptic function, coming from the transcriptomic analysis performed in mESCs subjected to E2f4 gene knock-out (KO) (see genes included in Supplementary Tables S1 and S5 from the study by Hsu and collaborators [19]). The transcriptional alterations in synaptic plasticity-related genes upon E2F4 modulation reveal the potential role of this protein in synaptic function. This suggests that E2F4 could be a promising target for several neurological diseases that course with synaptic plasticity impairment, such as AD.
E2F4 can interact with synaptic regulators. We verified using BP ontology that almost half of the E2F4 interactors found in the study by Hsu and collaborators [19], which are common in both mESCs and RPE-derived cells, have a function in either axonal transport or synapse physiology (Table S7). The E2F4 interactors involved in synaptic function that were identified in RPE-derived cells include Rac Family Small GTPase 1 (Rac1), cell division cycle 42 (Cdc42), and protein phosphatase 1 catalytic subunit β (PPP1CB) [19]. The actin regulators Rac1 and Cdc42 are important for the structural and functional plasticity of dendritic spines, which are the basis of learning mechanisms [101]. The actin cytoskeleton regulator Rac1 controls synaptic actin dynamics [102] and is involved in actin-regulated short-term presynaptic plasticity through the modulation of synaptic vesicle replenishment [103]. Cdc42 is known to have an important role in dendritic branching [104], and it is part of the mechanism involved in CaMKII activation, which modulates dendritic spine structural plasticity and induces LTP [105]. PPP1CB is one of the three catalytic subunits of protein phosphatase 1 (PP1), a serine/threonine protein phosphatase that regulates synaptic transmission and plasticity [106]. PP1 mediates NMDAR dephosphorylation, modulating the synaptic expression of this receptor [107]. Hsu and collaborators [19] also found Fragile X Mental Retardation Protein (FMRP) to be a candidate cofactor for E2F4 in mESCs. FMRP is an important regulator of activity-dependent plasticity in the brain, and the mutation in the FMR1 gene and subsequent loss of its protein product lead to Fragile X Syndrome (FXS), an inherited cause of autism and intellectual disability [108]. Mechanistically, FMRP is an RNA-binding protein that regulates the synthesis of synaptic and nuclear proteins within various compartments of the neuron [109]. FMRP binds to dendritic mRNA [110], and this may be important in mRNA localization to dendrites [111]. Thus, the hypothetical interaction of E2F4 with FMRP could be responsible for the modulation of synaptic protein transduction. Hsu and collaborators [19] also found that Snapin, a protein related to synaptic function [112,113], can interact with E2F4 in both mESC and RPE cells. In addition, the indirect effects of E2F4 on synaptic plasticity have also been described. In this regard, E2F4 can interact with Suv39H1 [114], a histone methyl transferase with an essential role in H3K9me3 methylation that mediates hippocampal memory functions [115]. The interaction of E2F4 with known synaptic regulators suggests that it may modulate synaptic function. This hypothesis is consistent with the observed enrichment of E2F1 in synaptic fractions, which is related to PSD95 expression and becomes upregulated with aging [116]. Furthermore, E2F1 is necessary for de novo neuronal tetraploidization occurring in mice, and this is associated with the alteration of cognition, as mice lacking this transcription factor show enhanced memory acquisition and consolidation [117]. Since E2F1 and E2F4 have antagonistic roles in neuronal function [96], we speculate that E2F4 could facilitate synaptic function and cognition, as opposed to E2F1.
Another piece of evidence for the putative capacity of E2F4 to regulate synaptic function comes from the study by [35], which showed that E2F4 can regulate genes involved in the MAPK signaling pathway. Although this pathway has been associated with cancer [35], it is also relevant for synaptic plasticity and AD [118,119,120]. A relevant member of the MAPK family of protein kinases is p38MAPK, the kinase that phosphorylates E2F4 in the Thr conserved motif controlling neuronal tetraploidization [13]. p38MAPK is a protein involved in synaptic plasticity and memory impairment that has been widely related to AD [120,121]. Accordingly, p38MAPK is progressively activated in neurons affected by AD [122] as well as in APP transgenic mice brains [121], and neuronal p38αMAPK mediates synaptic and cognitive dysfunction in a murine model of AD by controlling amyloid-β (Aβ) production [120]. Moreover, downregulation in APP/Tau-transgenic mice of p38MAPK results in the upregulation of genes involved in the MAPK pathway and calcium signaling [121]. Although the implication of E2F4 in this paradigm remains unclear, the expression of some calcium signaling and/or synaptic plasticity-related genes is altered upon p38α-MAPK deficiency in neuronal populations. In particular, the expression of both Grin2a and its encoded protein glutamate ionotropic receptor NMDA type subunit 2A (Grin2a) is decreased, resulting in a reduction of calcium influx in p38α-MAPK-deficient neurons [121]. Finally, knocking down E2f4 using an E2f4-specific shRNA significantly decreased the protein levels of p-ERK [35], a key MAPK that has been involved in both neurodegenerative diseases, as well as in endocannabinoid [123,124,125,126,127,128] and calcium signaling [101,105,129,130,131,132,133], which are critical pathways in synaptic function and modulation.
AD is a neurodegenerative condition that represents the most common form of dementia. It is characterized by memory and cognitive impairment, which are typically present in the early stages of the disease. Further clinical outcomes include a decline in visuo-spatial skills and neuropsychiatric disorders (apathy, irritability, aggressivity, wandering, and hallucinations). In a lower percentage, other AD symptoms include difficulty or inability to perform activities, olfactory disorders, pyramidal and extrapyramidal motor signs, myoclonus, seizures, and sleep complications [134,135]. AD is classified into two types, early onset AD (EOAD) or familial AD [136] and late-onset AD (LOAD) or sporadic AD [136]. From a neuropathological point of view, this disease is characterized by the presence of amyloid plaques, neurofibrillary tangles, neuroinflammation, and neurodegeneration in the brain [137]. AD is an unmet need, without any approved cure or disease-modifying therapy. Current treatments are addressed to ameliorate symptoms. Pharmacological treatments have evolved in recent years and have been based on drugs for neuropsychiatric symptoms, including antipsychotic, anxiolytic, anti-depressant and anti-convulsant drugs [138].
The etiology of AD is complex, and several hypotheses co-exist. The first descriptions of AD were based on the neuropathological phenotype of extracellular Aβ accumulation and neurofibrillary tangles, suggesting that Aβ processing was the upstream cause of AD [137,138]. Nevertheless, several studies support that Aβ processing abnormalities are necessary but not sufficient to lead to marked synaptic and neuronal loss [139]. Possible Aβ-independent mechanisms for AD etiology include synapse loss [140], altered glucose metabolism [141], cholesterol and lipid metabolism [138], oxidative stress [142], chronic hypoperfusion [143], cell adhesion pathways [138], immune system [138], and neuronal cell cycle re-entry [144] leading to tetraploidization [145]. The mutual interaction of all of these mechanisms makes it difficult to appropriately target the disease, and no effective therapies against AD are available until now. This is likely because the experimental therapies developed so far have mainly focused on single targets. Therefore, a paradigm shift is necessary, making it essential to design a multifactorial approach against this complex disease [146].
As indicated above, E2F4 can regulate more than 7000 genes involved in several activities key to AD progression, such as DNA repair, RNA processing, stress response, apoptosis, ubiquitination, protein transport and targeting, protein folding, and I-κB kinase/NF- κB cascade, according to studies performed in a lymphoblastoid cell line [27]. As observed in this cell line, as well as in mouse embryonic stem cells, E2F4 may activate or repress gene expression according to its interaction partner [19,27]. Interestingly, E2F4 malfunction has been linked to cognitive impairment [94], as well as to the etiopathology of neurodegenerative diseases, such as Alzheimer’s disease (AD) [11,12]. This is consistent with a recent study that proposes E2F4 as a major regulator of most AD-specific gene networks [95], and with other bioinformatics-based studies suggesting that E2F4 participates in this disease [147,148,149]. Moreover, a genome-wide association study for late-onset AD has identified a single nucleotide polymorphism that modifies a DNA-binding motif of E2F4 as relevant for the disease [150]. As mentioned above, E2F4 can be phosphorylated at multiple residues, including T248 (T249 in the mouse sequence) [19]. In vitro studies in differentiating chicken retinal neurons have provided insight that phosphorylation of E2F4 is key for the expression of cell cycle progression genes. In this model, nerve growth factor (NGF) can activate neurotrophin receptor p75, which in turn induces nuclear p38MAPK activity. As a result, E2F4 is phosphorylated in the T261/T263 motif, a change that allows cell cycle re-entry in these neurons [13], a mechanism generating neuronal tetraploidy [151]. We demonstrated in developing chick neurons that the expression of a dominant negative variant of chick E2F4 (E2F4DN) containing Ala substitutions in the Thr residues orthologous to T248 and T250 can prevent cell cycle re-entry in these cells and the subsequent DNA duplication that results in somatic neuronal tetraploidy [13]. Recent studies in our laboratory have confirmed that expression in the neurons of both mouse and human E2F4DN prevents neuronal tetraploidy in 5xFAD mice [12]. Moreover, as expected from a multifactorial factor, neuronal expression of E2F4DN was able to mitigate other processes that become affected in AD, such as neuroinflammation, Aβ peptide proteostasis, and body weight loss [12], a known somatic alteration associated with AD [152]. This results in the prevention of cognitive impairment [12]. Moreover, indirect evidence suggests that E2F4DN could also regulate synaptic function, as E2F4 has been shown to interact with a number of synaptic regulators in stem cells and in a photoreceptor-derived cell line (see above). All of these findings have allowed the development of an innovative gene therapeutic approach using human-derived E2F4DN [11] (see below). Based on the above evidence, we postulate that E2F4 represents a potential multifactorial target for AD, as this transcription factor possesses an intrinsic capacity to modulate several processes that are affected by this disease, thus reestablishing brain homeostasis and favoring brain tissue regeneration [19]. The homeostatic capacity of E2F4 could be crucial in counteracting any physiological stress [153] associated with the etiology of AD. In this context, the phosphorylation of E2F4 in the conserved T248/T250 motif could alter its homeostatic function, which would be restored by E2F4DN expression [11,12].
Human studies performed with AD brain samples have demonstrated that neurons overexpress cell cycle markers, including S-phase regulators [154,155,156,157,158,159,160]. This suggests that cell cycle re-entry participates in the etiology of AD. According to this hypothesis, a number of studies in mice have revealed that forced cell cycle re-entry in response to oncogene expression leads to the neuropathological hallmarks of AD, including Tau phosphorylation and neurofibrillary tangles [161,162], extracellular Aβ deposits [161], gliosis [163,164], synaptic dysfunction [165], neuronal death [163,165], and cognitive deficits [163], reinforcing that this process participates in the etiology of AD. Furthermore, neuronal cell cycle re-entry in humanized Aβ plaque producing mice results in the development of additional AD-related pathologies, namely, pathological tau, neuroinflammation, brain leukocyte infiltration, DNA damage response, and neurodegeneration [166]. Once neurons re-enter the cell cycle, DNA is duplicated, but neurons are rarely observed to undergo mitosis [145]. As a consequence, tetraploid neurons are generated [167,168,169], and this process represents an early hallmark of AD [170,171] that precedes [169,170] and recapitulates [169] the neuropathology associated with this disease. The increase in tetraploid neurons in the AD preclinical stage might contribute to cognitive impairment and neuronal death susceptibility at late stages [170]. Recent in vitro studies have shown that hyperploidy impacts neuronal morphology [172] and causes both synaptic dysfunction and delayed neuronal death [165], as previously observed in AD-affected hyperploid neurons [170]. A decreased density of PSD95 puncta and reduced AIS length correlated with an alteration in synaptic function and excitability in these neurons. Furthermore, neuron hyperploidization leads to diminished action potential generation and reduced spontaneous synaptic activity, with lower amplitudes of synaptic events when compared to control cells [165]. Interestingly, membrane depolarization with high K+, which mimics electrical input, increases the survival of hyperploid neurons without reversing synaptic dysfunction. Therefore, it has been hypothesized that AD-associated tetraploid neurons could be sustained in vivo if integrated into active neuronal circuits while promoting synaptic dysfunction. As a result of this synaptic dysfunction and enhanced survival, silent tetraploid neurons disturb the network of neural circuits, leading to the neurological abnormalities observed in AD. In fact, in silico studies have concluded that neuronal tetraploidy could lead to major effects in AD through alterations in the firing frequency caused by neuronal network disruption [172]. Therefore, the relationship between cell cycle reactivation and AD neuropathogenesis may rely, at least partially, on the generation of tetraploid neurons. Neuronal tetraploidy could also participate in the etiology of aging-dependent cognitive impairment, a process that takes place in individuals older than 40 years [173]. Indeed, a significant correlation between age and the proportion of tetraploid neurons was specifically observed in the entorhinal cortex of non-demented individuals [169], a known structure involved in memory formation [174]. In this context, age-associated neuronal tetraploidization can also be observed in the cerebral cortex of WT mice, while the blockade of neuronal tetraploidy in E2f1-deficient mice results in cognitive potentiation [169]. As indicated above, E2F4 controls cell cycle re-entry in neurons [13], and its expression becomes upregulated in cortical neurons from APP/PS1 mice [117]. A similar E2F4 upregulation is also observed in the prefrontal cortex of AD patients [144], as well as in neurons derived from human-induced pluripotent stem cells obtained from familial AD patients [148]. Furthermore, E2F4 becomes Thr phosphorylated in the cerebral cortex of APP/PS1 mice and Alzheimer’s patients [11,12,117]. As indicated above, phosphorylation of these two conserved Thr residues of E2F4 is necessary to induce neuronal tetraploidization and cognitive loss in AD, while expression of E2F4DN prevents these latter effects [11,12,117]. Therefore, E2F4 is a crucial agent regulating neuronal tetraploidization and its concomitant effects in the etiology of AD.
Studies performed in our laboratory have demonstrated that E2F4 fulfils a multifactorial effect in AD, as the expression of E2F4DN in neurons attenuates microgliosis and astrogliosis, two hallmarks of neuroinflammation, modulates Aβ peptide proteostasis and prevents body weight loss in 5xFAD mice. The paracrine effect of E2F4DN on neuroinflammation is likely mediated by either the cell membrane or extracellular factors released by E2F4DN-expressing neurons. Indeed, several mechanisms of bidirectional neuron–glia communication [175] have been described. In addition, neuron–glia communication can also take place through neuron-released exosomes [176]. The reduction of Aβ peptide levels in the hippocampus of 5xFAD mice in response to E2F4DN-based gene therapy [11] suggests the existence of a neuron-intrinsic capacity of the unphosphorylated form of E2F4 to prevent the production of this neurotoxic molecule. Nevertheless, the existence of a transcriptional program favoring Aβ peptide proteostasis in the double transgenic 5xFAD/E2F4DN mice suggests that E2F4DN may also induce cell-extrinsic effects on Aβ peptide proteostasis by acting on gene networks involved in processing, accumulation, and toxicity of Aβ [12]. E2F4DN expression in neurons can also reverse the loss of body weight observed in 5xFAD mice [11,12]. Since weight loss is likely associated with AD-associated metabolic alterations [177,178], the effect of E2F4DN on this trait may be due to a hypothetical capacity to affect neurons involved in sensing leptin [179], an adipocytokine that regulates energy metabolism and appetite [180]. E2F4 also has a connection with metabolic pathways since it can regulate insulin signaling in preadipocytes [74]. Therefore, E2F4 seems to participate in multiple pathways involved in energy metabolism and obesity, and this property may underline the capacity of E2F4DN to reverse weight loss in 5xFAD mice.
Neuronal E2F4DN expression prevents the cognitive deficits observed in 5xFAD mice [11,12], suggesting that E2F4 phosphorylation in the conserved Thr motif prevents its effects on multiple pathways involved in cognition, thus resulting in cognitive loss. Many regulatory pathways may favor cognitive rescue by E2F4DN. First, evidence has accumulated during the last decades connecting the cell cycle with synaptic plasticity, as common molecules are involved in both processes [181]. Therefore, the capacity of E2F4DN to prevent cell cycle re-entry in neurons and the concomitant tetraploidization process could prevent synaptic dysfunction in affected neurons [165]. Furthermore, the hypothetical capacity of E2F4 to regulate synaptic plasticity (see above) could also participate in the recovery of cognition observed in 5xFAD mice expressing neuronal E2F4DN [11,12]. Second, neuroinflammation has an important impact on synaptic plasticity and memory. On the one hand, activated microglia secrete cytokines, chemokines, and reactive oxygen species, which can lead to synaptic plasticity and memory deficits [182]. On the other hand, synapses can be functionally altered when astrocytes become reactive, thus causing hippocampal circuit dysfunction and memory alterations [183]. Therefore, the capacity of E2F4DN to attenuate neuroinflammation in 5xFAD mice may also account for its beneficial effects on cognition. Finally, the effects of neuronal E2F4DN expression on Aβ peptide proteostasis may favor cognitive recovery in 5xFAD mice, as this peptide is neurotoxic and known to trigger synaptic dysfunction and network disorganization [184].
To explore the therapeutic capacity of E2F4DN, we generated a knock-in (KI) mouse strain expressing mouse E2F4 with the T249A/T251A mutations (E2F4DN mice), Myc tagged at the C-terminus, and expressed under the control of the neuron-specific microtubule-associated protein tau (Mapt) promoter [12]. This transgenic mouse strain represents an optimal tool for the evaluation of E2F4 as a therapeutic target in neuropathology and brain aging. As a control, we used KI mice expressing EGFP under the Mapt promoter (EGFP mice) [185]. As mentioned above, hybrid mice resulting from the breeding of E2F4DN with 5xFAD mice (i.e., 5xFAD/E2F4DN mice) show a transcriptional program consistent with the attenuation of the immune response and brain homeostasis [12]. This correlates with the blocking of neuronal tetraploidization, the prevention of cognitive impairment, and the absence of body weight loss, a known somatic alteration associated with AD [152]. Consistently, 5xFAD/E2F4DN mice showed reduced microgliosis and astrogliosis at 3-6 months of age [12]. We further studied whether this effect is maintained at 1 year of age.
To verify whether microgliosis is reduced at 1 year of age in 5xFAD/E2F4DN mice compared with 5xFAD/EGFP mice, we crossed 5xFAD mice with either E2F4DN or control EGFP mice. Then, cortical sections of 1 year-old WT/EGFP, WT/E2F4DN, 5xFAD/EGFP and 5xFAD/E2F4DN mice were immunolabeled with Iba1, a specific microglia marker [186]. This analysis indicated that, as expected, the area occupied by microglia in the cerebral cortex of 5xFAD/EGFP mice was significantly greater than that of WT/EGFP mice (Figure 3a,b). This increase was associated with cortical layers 5–6 (Figure 3c). Therefore, as occurs at earlier time points [12], microglial cells are also activated in the cerebral cortex of 5xFAD/EGFP mice of 1 year of age. As observed at earlier time points [12], the presence of E2F4DN significantly diminished the area occupied by microglial cells in 1-year-old 5xFAD mice (Figure 3a,b), further supporting the hypothesis that neuronal E2F4DN attenuates the microgliosis observed in 5xFAD mice. Interestingly, E2F4DN was also able to prevent an increase in the area occupied by microglial cells in the cerebral cortex of WT/E2F4DN when compared with WT/EGFP mice (Figure 3a,b), confirming what was observed at 6 months of age [12]. These effects were observed in all cortical layers, except in layer 6, where WT/E2F4DN mice showed a non-significant tendency to decrease the Iba1-occupied area when compared with WT/EGFP mice (Figure 3c). Therefore, the previously described age-dependent increase in microgliosis in the cerebral cortex [187,188] is prevented by our therapeutic molecule. A significant reduction of the area occupied by microglia was also evident in the hippocampus of 1 year-old 5xFAD/E2F4DN mice when compared with 5xFAD/EGFP mice littermates of the same age (Figure 4). In addition, this same effect was observed in WT mice expressing E2F4DN (Figure 4), further supporting the hypothesis that the neuronal expression of our molecule can reverse the increase in microgliosis associated with brain aging in the hippocampus [187,188,189].
As mentioned above, reactive astrogliosis is known to increase with age in the cerebral cortex of 5xFAD mice compared to WT mice [12]. To study the effect of E2F4DN on the reactive astrogliosis observed in mature 5xFAD mice, we crossed 5xFAD mice with either E2F4DN or control EGFP mice. Then, cortical sections of 1 year-old WT/EGFP, WT/E2F4DN, 5xFAD/EGFP and 5xFAD/E2F4DN mice were immunolabeled with the specific actrocytic marker GFAP [190]. This analysis demonstrated that, as occurs with microglial cells, the area occupied by GFAP immunoreactivity in the cerebral cortex of 5xFAD/EGFP mice is significantly greater than that of WT/EGFP mice (Figure 5a,b). This increase was associated with all cortical layers except layer 1 (Figure 5c). Therefore, as occurs at earlier time points [12], microglial cells are also activated in the cerebral cortex of 5xFAD/EGFP mice of 1 year of age. As observed at 3 months [12], the presence of E2F4DN significantly diminished the area occupied by GFAP immunoreactivity in 1-year-old 5xFAD mice (Figure 5a,b), further supporting the hypothesis that neuronal E2F4DN expression attenuates the reactive astrogliosis observed in 5xFAD mice. This effect was observed in cortical layers 4 and 5 when 5xFAD/EGFP mice were compared with 5xFAD/E2F4DN (Figure 5c). This observation supports the hypothesis that the neuronal expression of E2F4DN can attenuate the increase in reactive astrocytes associated with AD. In contrast, E2F4DN was not able to reduce the area occupied by GFAP immunoreactivity in the cerebral cortex of WT/EGFP mice (Figure 5a,b), confirming what was observed at 3 months of age [12]. In the hippocampus, where GFAP was expressed by astrocytes at high basal levels in both WT and 5xFAD mice (Figure 6a), no difference was observed when E2F4DN was expressed in both WT/E2F4DN and 5xFAD/E2F4DN mice (Figure 6a,b).
In this review, we have included experimental results on two novel aspects of E2F4 function. On the one hand, by using an acetylated K96-specific antibody, we provide immunohistochemical evidence that E2F4 can be acetylated in K96 in neurons and cells located within the RMS, thus confirming the finding by Hsu and collaborators [19] demonstrating the presence of K96-acetylated E2F4 in mESCs and an RPE-derived cell line. On the other hand, we analyzed at 1 year of age the neuroinflammatory state of double transgenic 5xFAD mice expressing E2F4DN in neurons. This analysis constitutes a follow-up of a previously published study performed in transgenic mice of 3 and 6 months of age. Our results confirm the capacity of E2F4DN to attenuate microgliosis in the cerebral cortex and hippocampus of 5xFAD mice, even after one year, thus indicating that it has long-lasting therapeutic effects. In addition, E2F4DN was able to decrease the area occupied by GFAP cells (i.e., reactive astrocytes) in the cerebral cortex of the 5xFAD mice, while no changes in the area occupied by GFAP were observed in the hippocampus. This latter result contrasts with the observation that the area occupied by GFAP in the hippocampus of 3-month-old 5xFAD mice is decreased in the presence of neuronal expression of E2F4DN [12]. This discrepancy may be explained by the attenuation of astrocytosis in the hippocampus of 5xFAD mice at 1 year of age, a tissue where, in contrast to the cerebral cortex, GFAP is already expressed by non-reactive astrocytes. Therefore, E2F4DN-based gene therapy is likely to be effective in the long range. In this regard, we proved that our gene therapeutic approach is able to maintain the expression of E2F4DN for at least 1 year without major reduction in the transgene expression levels [11]. Since the durability of gene therapy has been reported to be good for years in humans [191], we expect that our E2F4DN-based gene therapy will require only one application for its effectiveness. Our results are also in favor of the hypothesis that E2F4DN plays a role in preventing brain aging. This is evidenced by the capacity of our therapeutic protein to reduce the levels of microgliosis in the cerebral cortex and hippocampus of 1-year-old WT mice, which is known to increase with age [187,188]. This result is consistent with a previous observation that the increase of microgliosis that is observed in the cerebral cortex of 6-month-old WT mice can be attenuated by the neuronal expression of E2F4DN [12]. Making an effort to better understand the non-canonical functions and mechanisms of action of E2F4 will greatly benefit many fields, including the study of neuronal function and malfunction associated with neurodegenerative diseases and brain aging. Although the role of E2F4 as a repressor of the cell cycle has been extensively studied, and its mechanism is fairly known, being a critical molecule in the RB/E2F pathway, little is known about E2F4 implications in other cell processes. The latest studies challenge this paradigm, indicating that E2F4 has several roles in cells in addition to this regulatory function in the cell cycle. In this review, we have discussed a new perspective focusing on the regulation by E2F4 of various biological programs in the cell, regardless of its classical function. We have discussed the possible mechanisms that support these new roles, as well as the implications of these functions for disease research, including neurodegenerative diseases, and brain aging. The potential versatility of E2F4 is intriguing, but given that E2F4 is broadly expressed in the cell, can modulate the expression of a wide variety of genes, and can bind to various targets, many of which are involved in fundamental neuronal processes, it makes sense to investigate non-canonical functions and to include E2F4 as a key protein in different cellular and, particularly, neuronal functions. Understanding these non-canonical functions will likely reveal new insights into its role in controlling neuronal activity and associated diseases, which in turn could guide the development of new strategies to treat neurodegenerative diseases and brain aging. Therefore, E2F4 is a potential therapeutic target for diseases with cognitive impairment, such as AD. As an example of the potentiality of E2F4 as an intervention target, we have discussed a novel mouse model expressing a mutant form of E2F4 that has proven to be a multifactorial therapeutic molecule for AD and likely for other neurodegenerative conditions and brain aging.
C57BL6/J mice were purchased from ENVIGO (Indianapolis, Indiana, USA). Double transgenic mice in C57BL/6J genetic background expressing mutant human APP695 containing the Swedish (K670N, M671L), Florida (I716V), and London (V717I) familial AD (FAD) mutations, and human presenilin 1 harboring the M146L and L286V FAD mutations, under the control of the Thy1 promoter (Tg6799 or 5xFAD mice) were purchased from the Jackson Laboratory (Bar Harbor, Maine, USA) (strain #008730). The 5xFAD mice were genotyped as indicated by the Jackson Laboratory. Homozygous Mapttm1(EGFP)Klt KI mice expressing enhanced green fluorescent protein (EGFP) in neurons (EGFP mice) [185] were purchased from The Jackson Laboratory (strain #004779). This mouse strain was maintained on a mixed background of C57BL/6 and 129Sv or backcrossed to the C57BL/6 background. EGFP mice have a target mutation in the Mapt gene, characterized by the insertion of the coding sequence of EGFP into the first exon, thus disrupting the expression of the tau protein. This results in the neuron-specific expression of cytoplasmic EGFP. Tau is expressed at high levels in neurons [192], and homozygous mice mutants for tau are viable, fertile, and display no gross morphological abnormalities in the central or peripheral nervous systems [185]. Homozygous EGFP mice are viable, fertile, normal in size, and do not display any gross physical or behavioral abnormalities. The EGFP mice were genotyped as indicated by the Jackson Laboratory. These mice were used in this study as a control for E2F4DN mice. Homozygous EGFP mice were bred with hemizygous 5xFAD mice to generate littermates consisting of hemizygous EGFP mice with or without the 5xFAD transgene. Mapttm(mE2F4DN-myc) KI mice (E2F4DN mice) were generated following the procedure described by [12]. These mice express a dominant negative form of E2F4 equivalent to the mutant E2F4 used to prevent NT in chick neurons [13]. The KI strain was maintained on a mixed background of C57BL/6 and 129Sv or backcrossed to the C57BL/6 background. Homozygous E2F4DN mice were created by inbreeding mice containing one copy of the E2F4DN transgene. Homozygous E2F4DN mice are viable, fertile, normal in size, and do not display any gross physical or behavioral abnormalities, even though the tau protein has been deleted [12]. Homozygous E2F4DN mice were bred with hemizygous 5xFAD mice to generate littermates consisting of hemizygous E2F4DN mice with or without the 5xFAD transgene. Analyses were performed on hemizygous mice for both Egfp and E2f4dn transgenes to avoid the observed effects of a full Mapt null mutation in the phenotype of APP and APP/PS1 transgenic mice [193,194,195]. E2F4DN mice are available upon request for research purposes other than neurological, neurodegenerative, and aging diseases.
The mouse anti-NeuN mAb clone A60 (MAB377; Merck Millipore, Burlington, Massachusetts, USA) was used at a 1:1000 dilution for immunohistochemistry. The rabbit anti-GFAP pAb (ab7260, Abcam, Cambridge, UK) was diluted to 1:1000 for immunohistochemistry. Rabbit anti-Iba1 pAb (019-19741, Wako) was used at a 1:800 dilution for immunohistochemistry. The anti-E2F4 (Acetyl-Lys96) rabbit pAb (D12062, Assaybiotech, Fremont, CA, USA) was diluted to 1:1,100 for immunohistochemistry. The donkey anti-rabbit IgG (H + L) highly cross-adsorbed secondary antibody Alexa Fluor 488 (Invitrogen, Waltham, MA, USA) was used at 1:1,000 dilution for immunohistochemistry. The goat anti-mouse IgG (H + L) cross-adsorbed secondary antibody Alexa Fluor 568 (Invitrogen, Waltham, MA, USA) was diluted 1:1000 for immunohistochemistry.
After anesthetizing the mice with intraperitoneal sodium pentobarbital (Dolethal; Vetoquinol, Alcobendas, Spain), administered at 50 mg/kg (body weight), they were transcardially perfused with phosphate buffered saline (PBS), and then with 4% paraformaldehyde (PFA). Brains were finally postfixed overnight at 4 °C with 4% PFA and cryoprotected by sinking in 32% sucrose in PBS at 4 °C. The brains were then embedded in 3% agarose gels prepared in 0.1 phosphate buffer, pH 7.37, before cutting them with a vibratome (50 μm). Vibratome sections were then stored at −20 °C in a solution of 3% glycerol (Panreac, San Fernando de Henares, Spain)/3% ethylene glycol (Panreac) prepared in 100 mM phosphate buffer, pH 7.37.
The vibratome sections were permeabilized and blocked in 0.4% Triton X-100 in PBS (PBTx) containing 10% fetal calf serum (FCS) for 3 h. They were then incubated overnight at 4 °C with the primary antibodies in 0.1% PBTx containing 1% FCS. After five washes of 20 min with 0.1% PBTx, the sections were incubated with the secondary antibodies plus 100 ng/mL DAPI in 0.1% PBTx for 3 h at room temperature. The sections were then washed five times with 0.1% PBTx, and mounted with ImmunoSelect antifading mounting medium DAPI (CliniSciences, Nanterre, France).
Lipofuscin was quenched with TrueBlackTM Lipofuscin Autofluorescence Quencher (Biotium, Fremont, CA, USA). Briefly, the vibratome sections were washed once with PBS and treated for 30 s with TrueBlack 1× prepared in 70% ethanol. Finally, the sections were washed three times with PBS, and then immunostained as described above.
Confocal images were acquired at 20× magnification with a Leica SP5 confocal microscope. Image analysis was performed using ImageJ (Fiji). The images used for the analysis (at least two mosaic images per tissue and animal) were maximum intensity projections created as output images whose pixels corresponded to the maximum value of each pixel position (in xy) across all stack images (z). DAPI staining was used to define the cortical layers and hippocampal structures. In order to analyze the area occupied by GFAP and Iba1, a threshold was set to highlight the area to be quantified. Quantification of the area occupied by Iba1-labeled microglia was achieved using a multi-step algorithm. First, Iba1-labeled microglia were segmented by applying a gray-scale attribute opening filter (area minimum: 25 pixels; connectivity: 8) to an 8-bit maximum projection. An opening morphological filter (1-pixel radius octagon) was then used effectively to separate microglia soma from processes before a maximum entropy threshold was used to discriminate microglial cells or astrocytes from the image background.
The quantitative data are represented as the mean ± s.e.m. Two-way ANOVA analysis was performed for the quantitative analysis of immune cells, followed by a post hoc Newman–Keuls test.
Gene ontology analyses (both CC and BP ontology) were performed using the database for annotation, visualization, and integrated discovery (DAVID) software [196,197] “https://david.ncifcrf.gov/ (accessed on 9 May 2022)”. | true | true | true |
|
PMC9603060 | 36165808 | Mazhar Hussain,Twyla Bradshaw,Morris Lee,Sassan Asgari | The Involvement of Atlastin in Dengue Virus and Wolbachia Infection in Aedes aegypti and Its Regulation by aae-miR-989 | 27-09-2022 | Aedes aegypti,dengue virus,atlastin,endoplasmic reticulum,microRNA,Wolbachia,mosquito | ABSTRACT Endoplasmic reticulum (ER)-shaping atlastin proteins (ATLs) have been demonstrated to play a functional role during flavivirus replication in mammalian cells. For dengue virus (DENV), atlastin is required in the formation of the replication organelles and RNA replication, virion assembly, production of the infectious virus particles, and trafficking or directing the association of vesicle packets with furin. Here, we investigated the involvement of atlastin in DENV replication in the mosquito Aedes aegypti and explored the possibility of its manipulation by the endosymbiotic bacterium Wolbachia to interfere with DENV replication. Results showed the expression of Ae. aegypti atlastin gene (AaATL) was upregulated in DENV-infected Aag2 cells, and its silencing led to reduced DENV replication. Contrary to our assumption that AaATL could be downregulated by Wolbachia, we did not find evidence for that in Wolbachia-infected cell lines, but this was the case in mosquitoes. Further, silencing AaATL did not have any effect on Wolbachia density. Our results also suggest that aae-miR-989 miRNA negatively regulates AaATL. The oversupply of the miRNA mimic led to reduced DENV replication consistent with the positive role of AaATL in DENV replication. Overall, the results favor AaATL’s involvement in DENV replication; however, there is no support that the protein is involved in Wolbachia-mediated DENV inhibition. In addition, the results contribute to discerning further possible overlapping functions of ATLs in mosquitoes and mammalian cells. IMPORTANCE Atlastin is a protein associated with the endoplasmic reticulum and has been shown to play a role in replication of flaviviruses in mammalian cells. This study aimed to investigate the role of mosquito Aedes aegypti atlastin (AaATL) in dengue virus replication and maintenance of Wolbachia, an endosymbiotic bacterium, in the mosquito. Our results suggest that AaATL facilitates dengue virus replication in mosquito cells, considering silencing the gene led to reductions in virus replication and virion production. Further, AaATL was found to be regulated by a mosquito microRNA, aae-miR-989. Despite an effect on dengue virus, AaATL silencing did not affect Wolbachia replication and maintenance in mosquito cells. The results shed light on the role of atlastins in mosquito-pathogen interactions and their overlapping roles in mosquito and mammalian cells. | The Involvement of Atlastin in Dengue Virus and Wolbachia Infection in Aedes aegypti and Its Regulation by aae-miR-989
Endoplasmic reticulum (ER)-shaping atlastin proteins (ATLs) have been demonstrated to play a functional role during flavivirus replication in mammalian cells. For dengue virus (DENV), atlastin is required in the formation of the replication organelles and RNA replication, virion assembly, production of the infectious virus particles, and trafficking or directing the association of vesicle packets with furin. Here, we investigated the involvement of atlastin in DENV replication in the mosquito Aedes aegypti and explored the possibility of its manipulation by the endosymbiotic bacterium Wolbachia to interfere with DENV replication. Results showed the expression of Ae. aegypti atlastin gene (AaATL) was upregulated in DENV-infected Aag2 cells, and its silencing led to reduced DENV replication. Contrary to our assumption that AaATL could be downregulated by Wolbachia, we did not find evidence for that in Wolbachia-infected cell lines, but this was the case in mosquitoes. Further, silencing AaATL did not have any effect on Wolbachia density. Our results also suggest that aae-miR-989 miRNA negatively regulates AaATL. The oversupply of the miRNA mimic led to reduced DENV replication consistent with the positive role of AaATL in DENV replication. Overall, the results favor AaATL’s involvement in DENV replication; however, there is no support that the protein is involved in Wolbachia-mediated DENV inhibition. In addition, the results contribute to discerning further possible overlapping functions of ATLs in mosquitoes and mammalian cells. IMPORTANCE Atlastin is a protein associated with the endoplasmic reticulum and has been shown to play a role in replication of flaviviruses in mammalian cells. This study aimed to investigate the role of mosquito Aedes aegypti atlastin (AaATL) in dengue virus replication and maintenance of Wolbachia, an endosymbiotic bacterium, in the mosquito. Our results suggest that AaATL facilitates dengue virus replication in mosquito cells, considering silencing the gene led to reductions in virus replication and virion production. Further, AaATL was found to be regulated by a mosquito microRNA, aae-miR-989. Despite an effect on dengue virus, AaATL silencing did not affect Wolbachia replication and maintenance in mosquito cells. The results shed light on the role of atlastins in mosquito-pathogen interactions and their overlapping roles in mosquito and mammalian cells.
Dengue virus (DENV) is transmitted by the primary mosquito vector Aedes aegypti that inhabits tropical and subtropical regions and is the leading cause of arthropod-borne viral diseases globally (1). Annually, there are about 390 million DENV infections, and 96 million of them are severe (2). DENV has an RNA genome of 10,700 nucleotides that includes 5′- and 3′- untranslated regions and a single open reading frame that encodes a single polyprotein (3). This polyprotein is cleaved into three structural proteins (capsid, premembrane/membrane, and envelope) and several nonstructural proteins, which are involved in viral RNA replication, virus assembly, and alteration of the host cell responses (4). Replication of DENV and other flaviviruses occurs on virus-induced host cell membranes and requires autophagy for efficient replication (3). There are four serotypes of DENV (DENV 1 to 4), each with genotype variability, and infection by one serotype results in exclusive long-term immunity to that serotype but not the others (5). Clinical symptoms of DENV in humans include dengue hemorrhagic fever/shock syndrome (DHS/DSS), severe muscle spasms and joint pain, and although most cases are acute, severe illness and death may occur (3). Vaccine development for DENV has been complicated due to serotype variability (6). The E protein, responsible for viral cell entry, has been the major target of vaccine development; however, E protein neutralization and protection across the four DENV serotypes has been limited (7). Upon human infection with one serotype, heterologous protection is short term, and previous infection with one serotype can predispose to DHS/DSS. Therefore, in areas of endemicity where more than one DENV serotype is circulating, a dengue vaccine faces the obstacles of risking an induced immunological condition, enhanced disease, and fleeting immunity (7). In the absence of an effective vaccine to protect against all the four serotypes or an antiviral drug, dengue disease control has centered around vector control or reduction in viral transmission using nonchemical approaches. One such approach is utilization of Wolbachia as a biological control agent (8). Wolbachia is a commonly found endosymbiotic bacterium in insects that depends on host nutrients such as amino acids and lipids, although it can be a nutritional symbiont (9, 10). It is mostly known for reproductive manipulations of the host, but Wolbachia has also been shown to stop replication of a variety of RNA viruses in Drosophila and mosquitoes; however, it is naturally less common in Ae. aegypti (10). Consequently, Ae. aegypti mosquitoes have been transinfected with a number of different strains of Wolbachia for population replacement with pathogen blocking property, and population suppression (8). However, the mechanism of pathogen blocking is poorly understood. The presence of Wolbachia induces changes in host gene expression and notable differences have been seen in antioxidant processes, metabolism, immune responses, and microRNAs (reviewed in references 11). In addition, Wolbachia competes with cytoplasmic replicating positive sense RNA viruses, such as DENV, for resources (11–13) and space (14). Therefore, a Wolbachia-modified host cellular environment is less favorable for cytoplasmic replicating viruses, such as DENV (15). The ER is the largest cellular organelle and functions in many cellular processes that are assimilated by flaviviruses once infected (16). It is composed of several subdomains, including perinuclear membrane sheets and tubule-like structures that branch through the cell periphery (17). Virus particles bud into the ER lumen upon entry into the cell and are translated at the rough ER. The expressed DENV proteins can induce rearrangement of ER membranes into three distinct structures in order to develop a suitable environment for viral replication (17, 18). These distinct structures include vesicle packets (VPs), which are suspected to be the site of viral replication and therefore, regarded as the viral replication organelle (RO). The other two rearrangements are convoluted membranes (CM), and membrane vesicles (VE) (17, 18). In Wolbachia-infected cells, it has been indicated that the ER could potentially be a source of vacuolar membrane and the interaction between Wolbachia cells (9). This particular organelle and derived intracellular vesicular trafficking play an important role in immune escape and control of apoptosis (9). Moreover, it has been shown that Wolbachia causes enhancement and redistribution of the ER at the subcellular level that results in highly enriched cytoplasmic regions of the tubular ER. Concurrently, a significant fraction of the ER contracts to become heavily clustered close to the nucleus (9). The ER subdomain of tubule-like structure extends through the cell periphery, and its morphology is maintained by distinct membrane shaping and fusion proteins (17). Such fusion proteins include the ER resident membrane-bound GTPases and atlastins (ATLs) (17, 19). ATLs are required to produce ER tubules in vitro, and their dysfunction leads to formation of long, unbranched ER tubules, and mutated ATLs have been associated with neurodegenerative diseases (17, 19). These ER-shaping responses suggest that ATLs are ER fusogens, which maintain branched ER tubule networks, key cellular factors in regulating ER function, and important factors in human disease (17, 19). ATLs in mammalian cells have been shown to be host factors exploited by both DENV and Zika virus (ZIKV), potentially influencing the DENV assembly or production of infectious particles and trafficking/directing the association of VPs with furin (17, 19). Considering the association of DENV and Wolbachia with the ER, we aimed to investigate the involvement of ATL in DENV and Wolbachia replication in the mosquito Ae. aegypti. While previous studies have indicated functional roles of ATLs in mammalian cells infected with DENV, to our knowledge, ATL has not yet been characterized for Ae. aegypti cells or challenged with Wolbachia and DENV. Previous studies of ATLs in mammalian cells determined that the proteins are required for the formation of the viral ROs and RNA replication, production of the infectious virus particles, and trafficking or directing the association of VPs with furin (17, 19). Moreover, current research has not yet deciphered the full mechanism(s) involved in Wolbachia’s virus-blocking phenotype. Therefore, we aimed to test the hypothesis that Ae. aegypti ATL (AaATL) has an involvement in flavivirus replication as seen in mammalian cells, and DENV replication is hindered upon Wolbachia infection because AaATL expression could be downregulated in the presence of Wolbachia. Further, we explored regulation of AaATL expression by a mosquito miRNA, aae-miR-989-3p. miRNA-target prediction suggested that among Ae. aegypti miRNAs aae-miR-989-3p has the highest potential to target AaATL. Further, this miRNA has been shown to be differentially expressed in response to Wolbachia and DENV infection in mosquitoes (20, 21).
When we searched the genome of Ae. aegypti, we could only find one atlastin gene (AaATL; NCBI XP_001663157.1, VectorBase AAEL003109), unlike human genome encoding three ATLs (ATL1-3); AaATL showed 57% and 59% amino acid sequence identities with HsATL2 and HsATL3, respectively (Fig. S1). Further, similar to HsATLs, AaATL contains a guanylate-binding domain (Fig. S1, the underlined region, aa 37 to 287). To assess the dynamics of the expression of AaATL upon DENV infection, Aag2 cells were infected with 1 MOI DENV and collected at 1 to 5 dpi. RT-qPCR analysis of RNA collected from cells showed gradual increase in AaATL transcript levels over time (Fig. 1A). The expression of AaATL at 5 dpi in comparison to that seen in mock-infected cells at 1 and 5 days was significantly higher (P = 0.0004), suggesting that AaATL could be involved during DENV replication in Aag2 cells. To ascertain DENV replication in the experiment, genomic RNA (gRNA) replication was assessed by RT-qPCR. As expected, mock cells showed no DENV replication. Over time, DENV replication could be seen to increase steadily (P = 0.001) following the days postinoculation and then plateaued at 5 dpi (Fig. 1B). The results suggested that Aag2 cells were successfully infected with DENV.
To silence AaATL, two synthetic siRNAs were used in transfection of Aag2 cells. AaATL expression was measured upon silencing the gene in transfected DENV-infected Aag2 cells (4 dpi). A very significant reduction (P = 0.0001) was observed in the expression of AaATL in cells transfected with AaATL siRNA 1 and 2 (Fig. 2A; 78% and 82% average reductions, respectively) compared to the control siRNA assessed through RT-qPCR. The result confirmed that AaATL was successfully silenced by both siRNAs. Considering both siRNAs were effective in silencing AaATL, for further experiments, we only used siRNA 1 (ATL siRNA). DENV genomic RNA replication was assessed at 2 and 5 dpi, after silencing AaATL in Aag2 cells with ATL siRNA to determine if replication of DENV was affected by atlastin expression. The RT-qPCR results showed significant reductions in DENV replication when AaATL was silenced at 2 dpi (P = 0.0051; Fig. 2B) and 5 dpi (P = 0.0008; Fig. 2C). Subsequently, focus-forming assay for DENV-2 virion titration was performed to verify the results of AaATL RNAi on DENV gRNA replication. Consistent with the gRNA results, at 2 and 5 dpi there were significant decreases (P = 0.0067 and P = 0.0265, respectively) in DENV virion titers (Fig. 2D and E). Further, silencing of AaATL in C6/36 cells, derived from Ae. albopictus, with siRNA led to significant reductions (P = 0.0006) in DENV virion production assessed by focus forming assay 3 dpi postinfection (Fig. 2F). AaATL and Ae. albopictus ATL are 90% identical at the nucleotide level with only one mismatch at the ATL siRNA site (Fig. S2). Overall, the results suggest that ATL could be involved in DENV replication in mosquito cells.
Since previous studies have shown that Wolbachia blocks DENV replication in mosquito cells, we investigated whether Wolbachia could reduce AaATL expression and consequently restrict DENV replication. For this, AaATL expression levels were assessed in two Aag2 cell lines persistently infected with two different strains of Wolbachia, wAlbB, and wMelPop. The expression levels were compared with that of the Aag2 uninfected cells. RT-qPCR analysis of RNA extracted from the cells suggested significantly higher AaATL transcript levels (P = 0.0008) in Aag2.wAlbB and Aag2.wMelPop cells compared to Aag2 cells (Fig. 3A). The effect of Wolbachia on AaATL expression was further evaluated in wAlbB-transinfected and tetracycline-cured female Ae. aegypti mosquitoes at various days post emergence (2 dpe, 6 dpe, 12 dpe). When the expression levels of AaATL between wAlbB and tetracycline-cured Ae. aegypti mosquitoes were compared by RT-qPCR, significantly less AaATL transcript levels were found in Wolbachia-infected mosquitoes at 2 dpe and 6 dpe (Fig. 3B). However, at 12 dpe, there were significantly higher AaATL transcript levels in Wolbachia-infected mosquitoes. These results in mosquitoes are somewhat contradictory to those seen in the cell lines (Fig. 3A) and suggest that change in AaATL expression could be age and/or tissue/cell specific. For example, while at 12 dpe there is more AaATL expression in Wolbachia-infected mosquitoes compared to Wolbachia-free mosquitoes, consistent with cell line results, at 2 and 6 dpi, there is less AaATL expression suggesting age-related regulation of AaATL. Aag2 cells are rather homogenous cells, whereas mosquitoes are comprised of different cell types and tissues that may have differential expression levels of AaATL. The density of Wolbachia was assessed upon RNAi of AaATL to evaluate whether the gene was required for Wolbachia maintenance in Aag2.wAlbB and Aag2.wMelPop cells. Silencing of AaATL was confirmed in the cells (Fig. 3C), however, there was no effect on Wolbachia density in either of the treated cells (Fig. 3D).
Expression of host miRNAs have been reported to be altered by flaviviruses infection (22, 23). For example, in Culex quinquefasciatus, miR-989 and miR-92 were differentially expressed in response to West Nile virus infection (24). In Ae. aegypti, 35 miRNAs were differentially expressed upon DENV-2 infection, four of which were upregulated and the rest downregulated (22). With the upregulation of AaATL over the course of DENV infection, we hypothesized that AaATL could be regulated by miRNAs. To investigate this, all the Ae. aegypti miRNAs were screened for potential binding sites to AaATL using RNA22. One of the top candidates turned out to be aae-miR-989 with very good seed region complementarity and minimum free energy of −22.10 kcal/mol (Fig. 4A). This target sequence is also conserved in Ae. albopictus atlastin, except the first nucleotide from the 5′ end (G/A). In Aag2 cells 5 dpi, we found high levels of AaATL transcript levels Fig. 1A. To find out if aae-miR-989 has any effect on AaATL transcript levels, Aag2 cells were transfected with aae-miR-989 mimic, and 24 h after transfection they were infected with DENV. Five dpi, there was a significant reduction (about 30%; P = 0.0011) in AaATL transcript levels compared to the controls transfected with negative-control mimic or the Cellfectin transfection reagent only (Fig. 4B). Increase in aae-miR-989 levels in mimic-transfected cells relative to NC was confirmed in the same samples (Fig. 4C). This suggested that the miRNA might negatively regulate AaATL expression. Next, replication of DENV in these cells was assessed. RT-qPCR results showed significant reduction (P = <0.0001) in DENV replication assessed by quantification of the viral gRNA (Fig. 4D). The titers of DENV virions also significantly declined (P = 0.0002) in the presence of aae-miR-989 mimic (Fig. 4E). This is consistent with the previous results that AaATL is involved in DENV replication; since aae-miR-989 reduces AaATL expression, in turn, this reduction has a negative effect on DENV replication. To further validate the sequence-specific interaction of aae-miR-989 with AaATL, the target sequence of AaATL was cloned downstream of the coding region for GFP in the pIZ/V5 plasmid vector. The vector was cotransfected into Sf9 cells together with aae-miR-989 mimic, NC mimic, or aae-miR-989 mimic with mutations in the seed region (Fig. 4A). While GFP transcript levels were significantly downregulated (P = 0.0004) in aae-miR-989 mimic transfected cells, NC or the mutant mimic had no effect on the GFP transcript levels (Fig. 4F). A similar result was observed at the protein level when cell samples from the same experiment were run on a Western blot, where anti-GFP was used as a control, although mutant mimic had some effect on GFP protein levels (Fig. 4G and H). Consistent with the results above, we found a negative correlation between the abundances of AaATL and aae-miR-989 in female mosquitoes from pupal stage to 4 days post adult emergence (Fig. 5A and B). Quantification of aae-miR-989 in samples from DENV-infected Aag2 cells (Fig. 1A) also showed a negative correlation between the miRNA and AaATL levels during DENV infection (Fig. 5C) and decrease in AaATL and increase in aae-miR-989 at 1 dpi followed by increases in AaATL levels from 2 to 5 dpi and reduced aae-miR-989 levels. Overall, the results suggested sequence-specific interaction of aae-miR-989 with AaATL target sequences and confirmed the negative regulation of AaATL by aae-miR-989.
Flaviviruses, such as DENV, replicate their genomes in altered ER membranes due to the organelle’s versatile functions in cellular processes (25). Since ATLs in mammalian cells have been recognized as membrane fusogens that maintain the ER membrane structure and facilitate DENV replication in the cells, the involvement of ATL during DENV replication and Wolbachia transinfection in the mosquito Ae. aegypti was investigated since it has not yet been characterized in mosquitoes. The effect and cross-effect of Wolbachia and DENV infection on AaATL was also investigated to test the hypothesis that DENV replication is hindered upon Wolbachia infection because the expression of AaATL is downregulated, and AaATL has a significant involvement in flavivirus replication as seen in mammalian cells. Flaviviruses use the ER membrane to establish their replication complexes (26). Atlastin enhances DENV replication by increasing the ER surface area. This is achieved by regulating ER tubule junctions by atlastin. Atlastins deficient cells lost their three-way ER tubule junctions (27). Further, in the characterization of human Atlastin 2, the GTPase activity was found to be important for DENV replication (17). Although human Atlastin 2 was found to interact with DENV NS3 protein, it is likely because of the regulation of ER tubule junctions by alastin. Both human Atlastin 2 and 3 affect DENV virion levels, but only Atlastin 2 reduces gRNA replication (17). Suppression of Atlastin 3 expression induced accumulation of immature DENV virions (17). Our results showed induction of AaATL following DENV infection in Aag2 cells and silencing the gene led to reduced DENV replication in Aag2 and C6/36 mosquito cells. These results suggest that similar to human atlastins, AaATL is a proviral factor for DENV. As there is only one atlastin in Ae. aegypti, it is possible that AaATL is involved in viral gRNA replication, virion assembly, and virion release. Future experimentation would still be ideal to further dissect the functional roles of AaATL, and this could involve examining virus protein localization. Therefore, an electron-microscopy analysis may increase understanding the function of AaATL. miRNAs have been shown to play a role in mosquito-arbovirus interactions, with several miRNAs found differentially expressed upon virus infection (28). However, only a limited number of target genes of these differentially expressed miRNAs have been identified. In this study, a miRNA, aae-miR-989-3p, was found to suppress the expression of AaATL. The expression of aae-miR-989 was upregulated in Zika virus–infected Ae. aegypti and DENV-infected C6/36 cells (21, 23). However, the role of aae-miR-989 in DENV or Zika virus infection is unknown. In this investigation, oversupply of aae-miR-989 reduced DENV gRNA and virion levels. In addition, the target site of aae-miR-989 was found to be conserved between Ae. aegypti and Ae. albopictus, consistent with silencing Ae. albopictus atlastin expression leading to reduced DENV replication in C6/36 cells. Therefore, aae-miR-989 may play a role in host response to DENV infection by inhibiting AaATL expression, although there could be more direct targets of aae-miR-989 that could contribute to DENV replication. Wolbachia is known to be associated with the ER and a recent study suggested a negative correlation between Wolbachia density and atlastin levels in Drosophila (29). In Wolbachia transinfected Aag2 cells, AaATL expression was upregulated in the case of Aag2.wAlbB and wMelPop cells; however, it is unknown whether this upregulation is caused by Wolbachia-induced ER remodeling (9). In Ae. aegypti mosquitoes transinfected with wAlbB, there was relatively less AaATL expression at 2 and 6 days post emergence but higher AaATL levels at 12 dpe compared to Wolbachia-free mosquitoes. Further, silencing AaATL in Aag2.wAlbB and wMelPop cells had no significant effect on Wolbachia density. The expression pattern of AaATL during Wolbachia or DENV infection did not support our hypothesis on AaATL being a factor in Wolbachia-mediated DENV inhibition. Overall, the results showed that consistent with human atlastins, AaATL is a proviral protein during DENV replication in Ae. aegypti cells. Further, we found that AaATL is negatively regulated by aae-miR-989 miRNA. Our results also showed that AaATL does not seem to be required for maintaining Wolbachia density or contribute to Wolbachia-meditated DENV inhibition in mosquito cells. The results shed further light on the tripartite Wolbachia-DENV-mosquito interactions.
The Ae. aegypti cell line (Aag2) persistently infected with Wolbachia strain wMelPop (Aag2.wMelPop) were used during in vitro experimentation and were grown in 1:1 Mitsushashi-Maramorosch and Schneider’s insect medium supplemented with 10% fetal bovine serum (FBS) at 27°C. In addition, Aag2.wAlbB cells, transinfected with purified wAlbB strain of Wolbachia from Aa23 cells, were used (30). Aag2, Aag2.wMelPop, and Aag2.wAlbB cells were infected at 1 MOI using ET300 DENV-2 stock prepared by amplification in C6/36 Aedes albopictus cells and diluted in 1:1 Mitsushashi-Maramorosch and Schneider’s medium containing 10% FBS. C6/36 cells were used for virus amplification because of an ineffective RNAi response during DENV replication due to a defective Dicer-2, which leads to higher virus yields (31). Mock-infected cells were used as control during the experiment and a 12-well plate was also prepared for infecting cells with medium containing 2% FBS. wAlbB-transinfected and tetracycline-cured Ae. aegypti mosquitoes (32) were used for analyzing AaATL expressions and RNA was extracted at 2, 6, and 12 days post-emergence (dpe).
Total RNA from cells and mosquito samples was extracted by using Qiazol, which was then DNase treated with TURBO DNase (Invitrogen). cDNA synthesis was performed by following the standard protocol of the M-MuLV reverse transcriptase kit (New England BioLabs) using specific primers for DENV-2 and oligo-dT primers for host genes (Table 1). A two-step qPCR was performed in duplicates using QuantiFast SYBR green PCR kit (Qiagen) and in a Qiagen Rotor-Gene Q under the following conditions: 95°C for 30s, and 40 cycles of 95°C for 10s and 60°C for 45s, followed by the melting curve (68°C to 95°C). For quantification of aae-miR-989, small RNAs were first reverse transcribed with miScript II RT kit (Qiagen) using the HiSpec buffer with 250 ng of total RNA per sample according to the manufacturer’s instructions. Quantitative PCR was followed with a miScript SYBR green PCR kit (Qiagen) in a Qiagen Rotor-Gene Q using 10 times dilution of cDNA per reaction. For aae-miR-989, the miRNA sequence was used for the forward primer sequence while 5s rRNA was used as the normalizing small RNA. RT-qPCR data were analyzed using the relative expression ratio method (Ratio = (Etarget)ΔCPtarget(control – sample)/(Eref)ΔCPref(control – sample)) as described previously (33). Gene expression levels or DENV gRNA levels in controls were adjusted to 1 and the transcript levels in treatments are expressed as fold changes relative to the controls.
The potential effect of AaATL on DENV replication in Aag2 cells and Wolbachia density in Wolbachia-infected cells was analyzed through RNA interference (RNAi)-mediated gene silencing of AaATL. Transfection in Aag2 cells was conducted with Cellfectin II (Invitrogen) and serum free transfection medium. Transfection was performed in 12-well plates in replicates, including Cellfectin II, a negative-control siRNA, or AaATL siRNA (synthesized by Genepharma) (Table 1). Cells were incubated overnight at 27°C and 24h after transfection they were infected with DENV at MOI 1. Cells and supernatants were collected at 2 and 5 days postinfection. AaATL was also silenced in Aag2.wAlbB and Aag2.wMelPop cells to assess effect on Wolbachia density by using the same siRNA transfection procedure mentioned above. For Wolbachia density, genomic DNA was first extracted from the transfected cells through the DIY spin column protocol (34). A two-step qPCR with melt curve analysis for Aag2.wAlbB and Aag2.wMelPop DNA was performed as per the manufacturer’s instructions (Qiagen) using previously developed primers that targeted the AeRPS17 gene and the Wolbachia surface protein gene (wsp) (Table 1).
Through focus forming assay, DENV-2 virions were titrated using C6/36 Ae. albopictus cells using the procedure previously described (35). Seeded in a 96-well plate, C6/36 cells were inoculated with serially diluted medium (10−1, 10−2, 10−3, 10−4) collected from experiments. Plates were rocked for 1 h at room temperature and then incubated at 37°C for another hour. After incubation, the medium was removed, and cells were overlaid with 1.5% carboxymethyl cellulose (CMC) and 2.5% FBS in Opti-MEM medium. A following incubation period of 72 h at 27°C occurred before removing the overlay and fixing the cells for 20 min at −20°C with 80% ice-cold acetone in PBS before air drying overnight. PBST was used to block the cells at 37°C for 30 min before undergoing a 2 h of incubation period at 37°C with the primary DENV-2 envelope (human) antibody diluted (1:1000) with PBST. Plates were washed thrice with PBST and incubated for 1 h at 37°C with the secondary antibody (IRDye800CW goat anti-human LICOR). A second washing of the plates occurred, followed by a drying and scan on the Odyssey imager (LI-COR Biosciences) at 41 mM resolution to count foci and calculate viral titer, respectively. Focus forming units (FFUs) were obtained by performing titrations in triplicates.
To find potential target site(s), the full-length sequence of AaATL mRNA was used against all the known Ae. aegypti miRNAs available on miRBase. For this, we used RNA22 software (https://cm.jefferson.edu/rna22/Interactive/) given stringent criteria, including maximum complementarity in the seed region (nucleotides 2 to 8 from the miRNA 5′ end) as well as maximum folding energy for heteroduplex (Kcal/mol). aae-miR-989 showed the best folding energy (−22.10 Kcal/mol) and complete complementarity with the seed region. To explore the interaction of aae-miR-989 with AaATL as well as effect on DENV replication, Aag2 cells were transfected with aae-miR-989 mimic and negative-control mimic (synthesized by Genepharma) 100 μM each, or Cellfectin II transfection reagent only. After 24 h of transfection, cells were infected with DENV at MOI 1 and collected at 5 dpi. To assess the direct interaction of aae-miR-989 with AaATL target sequences, two oligonucleotides (42 nt each Forward and Reverse, Table 1) containing target sequence were designed with restriction sites of XbaI and SacII incorporated at the 5′ and 3′ ends, respectively. The sequence of reverse oligonucleotide was reverse complementary to the forward one. For annealing, both oligonucleotides (10 μM each) were mixed and annealed at 94°C for 5 min and slowly cooled down to room temperature for 45 min. The resulting dimer was digested with XbaI and SacII restriction enzymes, purified and cloned into the pIZ/GFP vector downstream of the GFP coding region. Clones were sequenced for confirmation. The plasmid (1 μg) and 100 μM mimic, control mimic, mutated mimic, or Cellfectin II transfection reagent only were cotransfected into Sf9 cells, which were grown in SF900 III medium (ThermoFisher Scientific). We used Sf9 cells as a heterologous cell line to assess the direct interaction of aae-miR-989 with AaATL target sequences. Heterologous cell lines are commonly used for miRNA-target interaction experiments with a reporter gene to avoid the effect of the endogenous miRNA being studied on the reporter construct. A miRBase (mirbase.org) search of Spodoptera frugiperda, from which Sf9 cells are derived, showed no match for aae-miR-989. Three days posttransfection cells were collected from which RNA was extracted and subjected to RT-qPCR using specific primers to GFP (Table 1) to assess the expression levels of GFP. Protein samples from the same experiment as above were also assessed on a Western blot using an anti-GFP antibody (Invitrogen) as the primary antibody and an anti-rabbit (Sigma) antibody as the secondary antibody both at a 1:10,000 dilution. For reference protein, the same blot was hybridized with anti-Histone 3 (Life Technologies) followed by secondary anti-mouse antibody (Invitrogen) both at 1;10,000 dilutions. The blot was developed by using nitroblue tetrazolium chloride (NBT) and 5-bromo-4-chloro-3-indolylphosphate (BCIP) reagents. The density of GFP bands were semiquantified relative to those of the histone protein using ImageJ. | true | true | true |
PMC9603156 | Yu-Jen Chen,Chia-Tien Hsu,Shang-Feng Tsai,Cheng-Hsu Chen | Association between Circulating MicroRNAs (miR-21-5p, miR-20a-5p, miR-29b-3p, miR-126-3p and miR-101-3p) and Chronic Allograft Dysfunction in Renal Transplant Recipients | 14-10-2022 | chronic allograft dysfunction,microRNA,biomarker | Chronic allograft dysfunction (CAD) is a major condition affecting long-term kidney graft survival. Serum microRNA (miRNA) has been reported as a biomarker for various conditions of allograft injuries. The upregulation of miR-21 is the best-known miRNA change in graft tissue, urine and plasma. However, the correlation of plasma miR-21 with the severity of CAD remains unclear. In our study, 40 kidney transplantation recipients with late graft survival for more than 10 years were enrolled. The CAD group (n = 20) had either an eGFR between 15 to 60 mL/min or a biopsy-proved chronic allograft nephropathy or rejection. The control group (n = 20) had an eGFR ≥ 60 mL/min without proteinuria and hematuria for a consecutive 3 months before the study. We performed RNA sequencing to profile the miRNAs expression. There were six differentially expressed miRNAs in the CAD group. Among them, miR-21-5p and miR-101-3p were decreased, and miR-20a-5p was increased. We found that miR-21-5p, miR-20a-5p and miR-101-3p all participated in the TGF-beta pathway. We demonstrated that decreased miR-21-5p and miR-101-3p, and increased miR-20a-5p were the novel CAD-associated miRNAs in the TGF-beta pathway. These findings may pave the way for developing early prediction miRNAs biomarkers for CAD, and possibly developing therapeutic tools in the field of kidney transplantation. | Association between Circulating MicroRNAs (miR-21-5p, miR-20a-5p, miR-29b-3p, miR-126-3p and miR-101-3p) and Chronic Allograft Dysfunction in Renal Transplant Recipients
Chronic allograft dysfunction (CAD) is a major condition affecting long-term kidney graft survival. Serum microRNA (miRNA) has been reported as a biomarker for various conditions of allograft injuries. The upregulation of miR-21 is the best-known miRNA change in graft tissue, urine and plasma. However, the correlation of plasma miR-21 with the severity of CAD remains unclear. In our study, 40 kidney transplantation recipients with late graft survival for more than 10 years were enrolled. The CAD group (n = 20) had either an eGFR between 15 to 60 mL/min or a biopsy-proved chronic allograft nephropathy or rejection. The control group (n = 20) had an eGFR ≥ 60 mL/min without proteinuria and hematuria for a consecutive 3 months before the study. We performed RNA sequencing to profile the miRNAs expression. There were six differentially expressed miRNAs in the CAD group. Among them, miR-21-5p and miR-101-3p were decreased, and miR-20a-5p was increased. We found that miR-21-5p, miR-20a-5p and miR-101-3p all participated in the TGF-beta pathway. We demonstrated that decreased miR-21-5p and miR-101-3p, and increased miR-20a-5p were the novel CAD-associated miRNAs in the TGF-beta pathway. These findings may pave the way for developing early prediction miRNAs biomarkers for CAD, and possibly developing therapeutic tools in the field of kidney transplantation.
Potent immunosuppressive agents markedly improve renal graft survival rates. The 5-year graft survival rate of deceased transplant donors (DTD) combined with living transplant donors (LTD) was 88.7% in Taiwan during the years 2009 to 2013 [1]. Meanwhile, in the USA during the period 2012 to 2015, the graft survival rate of LTD was 93.29% and that of DTD was 78.15% [2]. Maintaining long-term graft survival remains a major unmet need. The 10-year patient graft survival rate of deceased kidney transplant recipients was 53.62%, and that of living kidney transplants was 81.28% in the USA from 2008 to 2011 [2]. The leading etiology of kidney graft failure includes alloimmune injury and recurrent glomerulonephritis [2]. Although a kidney biopsy can provide valuable information on the differential diagnosis of allograft dysfunction, it is an invasive procedure and is not always readily available. Chronic allograft dysfunction (CAD) is defined as a clinical condition characterized by a slowly progressive drop in kidney function, usually associated with hypertension and proteinuria. The etiology of CAD includes rejection, BK polyomavirus nephropathy, graft renal artery stenosis, and calcineurin inhibitor (CNI) nephropathy. MicroRNAs (miRNAs) are small non-coding RNAs with lengths ranging from 18 to 24 nucleotides. They are stable in tissues and body fluids, and are promising, non-invasive biomarkers for disease diagnosis and forensic science [3,4]. Additionally, they are preserved in eukaryotic organisms and regulate the post-transcriptional expression of genes, while providing insight into the molecular pathway of different renal injuries following kidney transplantation. Previous kidney allograft studies have focused on acute kidney injury, acute or chronic rejection and interstitial fibrosis/tubular atrophy (IF/TA). Kidney fibrosis is observed in the graft histology of CAD patients, and the TGF-β pathway is well known for causing tissue fibrosis. Several microRNAs have been reported to regulate the TGF-β pathway in animal models and human studies [5]. In samples of graft biopsies revealing IF/TA, their miR-21 levels had increased [6,7]. Additional studies have also suggested that both plasma miR-21 [7,8,9] and urinary miR-21 [10,11] increased in IF/TA. Other miRNAs, including miR-142-3p, miR-142-5p, miR-155, miR-200b and miR-29 are also related to IF/TA [6,9,12]. Here, we aimed to identify miRNAs signatures in long-term CAD and also explore any potential targets for diagnosis and treatment.
Demographic data surrounding the two groups are shown in Table 1. No inter-group differences were found regarding age, graft survival, donor status or malignancy after transplantation. There were more females in the CAD group compared with the control group (12 vs. 5, p = 0.025), while the FK-506 level was higher in the CAD group (5.7 ± 0.9 vs. 5.1 ± 1.2, p = 0.829). Additionally, there were apparently fewer cases of malignancy in the CAD group when compared with the control group (2 vs. 5, p = 0.407). In the CAD group, the histology of renal specimens revealed five mildly acute T-cell rejections, two chronic antibody-mediated rejections (CAMRs), two tubular injuries and one striped fibrosis. In the control group, two patients had undergone a biopsy 10 years after kidney transplantation, and both of their histology revealed acute tubular injury. Of the eighty-nine miRNAs in the CAD group, we found eight to be differentially expressed (i.e., miR-21-5p, miR-15a-5, miR-101-3p, miR-589-5p, miR-122-5p, miR-20a-5p, miR-29b-3p and miR-126-3p). Amongst them, three were up-regulated (miR-20a-5p, miR-589-5p and miR-29b-3p) and five were down-regulated (miR-21-5p, miR-15a-5, miR-101-3p, miR-122-5p and miR-126-3p) in the CAD group. The miRNAs were identified in accordance with our criteria of fold change (≥±0.585) and p values (<0.05) as shown in Table 2 and Figure 1. The clustering analysis and PCA plot can be found in Supplementary Figures S1 and S2. In silico analysis on the targeted genes of the eight differentially expressed miRNAs was performed, and we found miRNA target interaction (MTI) in miRTarBase [14]. The top ten GO enrichments of the biological process (BP) are listed in Table 3. Extracellular matrix organization, extracellular structure organization and external encapsulating structure organization were three of the top four terms according to their p-value. Other GO terms are listed in Supplementary Tables S1 and S2.
In this case-control study, we explored the miRNAs signature of CAD with a graft survival period of more than 10 years. Of the eight differentially expressed miRNAs, miR-21-5p, miR-15a-5p, miR-101-3p, miR-122-5p and miR-126-3p were detected in more than nineteen samples in both the CAD and control groups. Enrichment analyses revealed that miR-21-5p, miR-20a-5p, miR-101-3p, miR-126-3p and miR-29b-3p participated in the TGF-B/Smad pathway. miR-21 plays a crucial role in development, cancer, cardiovascular kidney disease, the aging process and inflammation [15]. Ghorbanmehr et al. found that urinary miR-21-5p could be differentially detected in prostate and bladder cancer patients [16]. A growing body of literature suggests that miR-21 plays an important role in chronic kidney disease in animal models and human studies [17,18]. miR-21 was upregulated in the TGF-β1/Smad pathway, as reported in several studies [19,20]. TGF-β is known for its anti-inflammatory, anti-neoplasm and fibrosis functions. The TGF-β canonical pathway is illustrated and described in Figure 2, however the non-canonical pathway was beyond the scope of our study. The association between miR-21 and fibrosis can be explained by its effect on the down-regulation of Smad7, which inhibits Smad3 [19,20]. Zang et al. found elevated urinary exosome mir-21-5p in diabetic kidney disease patients [21]. Elevated levels of miR-21 in human graft biopsy tissue were found to be associated with tubulointerstitial fibrosis [6], while elevated urinary miR-21 has also been associated with elevated levels of miR-21 in graft tissue [10,11,12]. However, a similar correlation with plasma miR-21 levels is less clear. Glowacki et al. reported that plasma miR-21 elevation is associated with severe IF/TA in graft kidney [7]. Alternatively, Saejong et al., reported that whole-plasma miR-21 is lower in CAD cases, a finding that is consistent with our present results. Interestingly, the plasma exosome miR-21 level was higher in CAD, thus making it better correlated with the severity of IF/TA [8]. miR-126-3p was reported to be associated with malignancy [22,23,24], particularly in lung and breast cancer patients [25,26,27,28,29]. Jordan NP et al. found that miR-126-3p was down-regulated in the fibrotic tissue of murine kidney and heart, with the process being due to the endothelial-to-mesenchymal transition [30]. Motshwari et al. found increased whole-blood miR-126-3p in a South African community-based sample of subjects diagnosed with chronic kidney disease as compared to a control group [31]. Although there was no kidney transplant study performed, Schaefer et al. demonstrated that pre-treatment with S-NO-HSA led to reduced fibrosis and the preservation of myocardial miR-126-3p and GATA2 levels in murine cardiac isografts 60 days after transplantation [32]. Therefore, miR-126-3p may serve as a potential therapeutic target in the organ transplantation field. miR-101-3p has been reported on over the last 3 years. Wang et al. found miR-101 reverse TGFβ1 induced epithelial-to-mesenchymal transition (EMT) in HK 2 cells with less fibrosis [33]. Zhao et al. also revealed that miR-101 inhibits AKI-to-CKD transition by regulating EMT [34]. One additional study also demonstrated that miR-101 suppresses chronic renal fibrosis by regulating KDM3A, hence blocking the YAP-TGF-β/Smad signaling pathway [35]. Interestingly, one study showed that dexmedetomidine potentially protects against renal fibrosis by targeting the miR-101/TGF-β/Smad pathway. miR-101-3p may serve as a promising diagnostic marker of CAD. However, further studies surrounding histology, urinary and plasma expression in patients with kidney transplants are still required. miR-20a-5p can repress myoblast proliferation in avian cells [36]. It was also reported to be upregulated during the first few days after acute ischemic reperfusion injury in a mouse model [37]. Its correlation with fibrosis in kidney disease was not made clear, but miR-20a-5p was down-regulated in liver fibrotic tissue [38]. Smad4 was the target of miR-20 in the pathway [39]. Interestingly, several studies have suggested that miR-20a-5p could be a biomarker for malignancy due to Smad4 serving as a co-Smad in the TGF-β pathway [40,41,42,43]. We also observed fewer malignancy cases in the CAD group. In the kidney disease studies, Smad4 deficiency in a unilateral ureteral obstruction (UUO) model likely enhances both renal inflammation and fibrotic response [20]. To the best of our knowledge, the association between miR-20 and human chronic kidney disease or allograft injury has never been reported. The miR-29 family has been well known for their inhibition of extracellular matrix-related genes, including COL1A1, COL3A1, COL4A1, MMP2, TGFB1 and TGFB2 [5,44,45,46,47] for more than 10 years, while also having the most potential as a therapeutic target for antifibrosis treatment [48]. Zhang et al. found that long noncoding RNA Tug1 promotes angiotensin-II-induced renal fibrosis by interacting with miR-29b-3p [49]. Wang et al. demonstrated that exosome-mediated miR-29 transfer could reduce muscle atrophy and kidney fibrosis UUO in a mice model [50]. Zhang et al. also found that rAAV9-mediated supplementation could improve angiotensin-II-induced renal fibrosis in mice [51]. Gondaliya et al. also demonstrated that miR-29b could attenuate histone deacetylase-4-mediated podocyte dysfunction and renal fibrosis in diabetic nephropathy [52]. Although miR-29 is upregulated in fibrotic tissue, serum level has not frequently been reported to be correlated with fibrosis. Rubis et al. found decreased miR-21-5p and elevated miR-29b in biopsy-proven fibrotic dilated cardiomyopathy patients [53]. Therefore, miR-29b-3p could be a promising therapeutic target for allograft dysfunction. As for miR-15a-5p and miR-122-5p, several studies have been published regarding fibrosis in different organs [54,55,56,57]. Liu et al. compared different treatments involving rAAV-miR-122-5p and rAAV-GFP in 14-week-old male SHR and WKY rats. The team demonstrated that miR-122-5p possesses potential therapeutic significance for hypertensive renal injury and fibrosis-related kidney diseases [58]. Dieter et al. examined the urine and plasma levels of miR-15a-5p and miR-30e-5p in patients with type I DM and CKD., with no difference being found regarding miR-15a-5p levels [59]. The above two miRNAs still require more studies to be performed in order to confirm their roles in human kidney disease. Multiple pieces of evidence surrounding elevated miR-21 being in association with renal fibrosis have been reported. However, in our patients, we observed a lower plasma miR-21-5p level compared with the control group. Li et al. found elevated circulating miR-21-5p after steroid treatment had been performed in humans and rats [60], possibly indicating that anti-inflammatory status may be correlated with elevated circulating miR-21-5p. We postulated that miR-21-5p, miR-20a-5p and miR-29b-3p serve as a housekeeping function, are secreted by various cells and can be detected at a baseline level. When allograft dysfunction gradually developed, allograft kidney increased the uptake of miR-21-5p and decreased the uptake of miR-29b-3p and miR-20a-5p. A serial follow up of multiple microRNAs could be utilized as a surveillance or prediction tool for allograft dysfunction. Hromadnikova et al. were able to predict that 47.93% of patients were destined to develop gestational diabetes mellitus (GDM) at a 10.0% false-positive rate during the first trimester of pregnancy, with 11 dysregulated microRNAs. After incorporating clinical characteristics, the team increased their detection rate to 72.5%, with a 10% false-positive rate [61]. We will consider combining the microRNAs together and then evaluating their performance when predicting CAD. Our study had the limitation of potential bias due to its retrospective nature and small sample size. Additionally, we did not implement the protocol-directed biopsy in our hospital. Consequently, a heterogeneity of diagnosis likely existed in the allograft dysfunction group, and therefore the interpretation of the results should be prudent. In conclusion, the plasma microRNA signature of downregulated miR-21-5p, miR-101-3p and miR-126-3p, as well as upregulated miR-20a-5p and miR-29b-3p is a promising biomarker for detecting CAD. MiR-29b-3p offers great potential towards attenuating chronic allograft dysfunction.
Blood samples were collected, and plasma was extracted within 2 h of admittance according to standard procedures, with the samples first being centrifuged at 1200× g for 10 min, and the supernatant again being centrifuged at 12,000× g for 10 min. The supernatant, which was plasma, was divided into aliquots and immediately stored at −80 °C. The miRNA was isolated from 200 μL of plasma according to the miRNeasy manufacturer’s protocol. Plasma miRNAs were eluted in 20 μL of nuclease-free water. Concentrations of the extracted miRNAs were quantified using the Qubit microRNA Assay Kit (Q32880, ThermoFisher Scientific Inc., Waltham, MA, USA).
From the samples, 2 ng of the total miRNAs were used to synthesize cDNA with 20 μL reverse transcription reactions. The reverse transcription step was performed as follows: poly-A tail was added to the miRNA population using Poly-A polymerase, followed by cDNA synthesis with a QuarkBio’s microRNA Reverse Transcription Kit (Quark Biotechnology, Inc., Zhubei, Taiwan). The qPCR was performed utilizing the NextAmp™ Analysis System and mirSCAN™ V2 PanelChip®. For qPCR analysis, 0.15 ng cDNA was added to the QuarkBio qPCR master mix (Quark Biotechnology, Inc., Zhubei, Taiwan), and Q Station™ (Quark Biotechnology, Inc., Zhubei, Taiwan) was run according to the following cycle program: 95 °C for 36 s and 60 °C for 72 s over 40 cycles. For miRNA expression profile analysis, SYBR Green-based qPCR methodology used the miScriptmiRNA PCR Array Human T-Cell and B-Cell Activation panel (Qiagen, Las Matas, MD, Spain), allowing for the quantification of 89 miRNAs. This method was used to identify those miRNAs differentially expressed (DE) in 20 transplant samples with CAD and 20 transplant samples having stable renal function. RT-PCR data were normalized according to the manufacturer’s instructions. The miRNAs were selected based on the fold changes of ≥±0.585 and at p values < 0.05.
Once the differentially expressed miRNAs were detected, we determined miRNA target interaction (MTI) using miRTarBase. We filtered out the MTIs with less than 3 reference supports as well as any non-functional MTIs. Genes listed within MTIs were then analyzed for gene set enrichment using clusterProfiler. Gene Ontology, KEGG Pathway and Disease Ontology were used for functional analysis due to their long-standing curation.
We performed gene set enrichment analysis on Pathway terms and Gene Ontology (GO) terms with the input of target genes from differentially expressed miRNAs. Overrepresented gene sets were performed using GOSemSim in order to calculate the similarity of GO terms and remove the highly similar terms by keeping one representative term. A background gene set was based on validated miRNA target interactions indicated by the miRTarbase (2766 genes).
Principal component analysis (PCA) was performed in order to evaluate the differences between biological replicates and their treatment conditions. PCA adopted an orthogonal transformation to convert a set of observations of possibly correlated variables into a set of uncorrelated variables called principal components.
For advanced data analyses, intensity data were pooled and calculated to identify differentially expressed microRNAs based on the threshold of fold changes and p-values. The correlation of expression profiles between samples and treatment conditions was demonstrated through unsupervised hierarchical clustering analysis.
Continuous variables were summarized as mean ± standard error or median with interquartile range. Continuous variables with normal distribution were compared using Student’s t-test or Fisher’s exact test, and those with non-normal distribution were compared with the Mann–Whitney U test. Categorical variables were compared using the χ2 test or Fisher’s exact test. All statistical tests were 2-tailed, and a p value < 0.05 was considered statistically significant. Statistical analyses were conducted using IBM SPSS Statistics 26 (SPSS Inc., Chicago, IL, USA). Plots were generated using Prism 6 (GraphPad, San Diego, CA, USA).
We retrospectively recruited a total of 40 kidney transplant recipients who were experiencing late renal graft survival for a period of more than 10 years from Taichung Veteran General Hospital. We recorded their age, gender, graft survival, donor status, cause of end stage renal disease and malignancy after transplantation. All blood samples were collected after receiving their informed consent. Patient data included estimated glomerular filtration fraction (eGFR), plasma creatinine concentration (SCr), urine protein creatinine ratio, calcineurin inhibitor dosage, calcineurin inhibitor plasma level, prednisolone daily dosage and mycophenolic acid daily dosage. The eGFR was calculated from plasma creatine using the Modification of Diet in Renal Disease (MDRD) equation. Recipients in the CAD group (n = 20) included patients with their eGFR between 15 and 60 mL/min, or those who had biopsy-proved chronic transplant allograft nephropathy or rejection. The control recipients (n = 20) had an eGFR ≥ 60 mL/min without either proteinuria or hematuria for 3 consecutive months prior to the study. We performed cDNA synthesis and RT-PCR on the plasma of each patient in order to profile miRNA expressions and predict the potential targets of differentially expressed miRNAs in CAD. There were 10 patients in the CAD group and 2 patients in the control group who had previously received a kidney allograft biopsy prior to blood sampling. | true | true | true |
|
PMC9603196 | Veronica Dimuccio,Linda Bellucci,Marianna Genta,Cristina Grange,Maria Felice Brizzi,Maddalena Gili,Sara Gallo,Maria Laura Centomo,Federica Collino,Benedetta Bussolati | Upregulation of miR145 and miR126 in EVs from Renal Cells Undergoing EMT and Urine of Diabetic Nephropathy Patients | 11-10-2022 | diabetic nephropathy,urinary extracellular vesicles,miRNAs,epithelial to mesenchymal transition,biomarkers | Simple Summary Diabetic nephropathy is one of the most frequent complications of diabetes, resulting from diffuse damage to different kidney cells. The identification of subjects at risk is mandatory to prevent its development and provide appropriate therapies reducing the unmanageable evolution towards end-stage kidney disease. The aim of this work was to identify urinary-derived extracellular vesicles (EVs) miRNA cargo to be used as biomarker of kidney damage in diabetic patients. The miRNA profile was then correlated with the molecular mechanism associated with the glomerular and tubular damage using a diabetic-like model. In patients, miR145 and miR126 in urinary EVs increased together with albuminuria. MiR145 and miR126 increased in parallel in EVs from renal epithelial cells undergoing transition to a fibrotic mesenchymal phenotype. These data unveiled a role for miR126 and miR145 as the biomarkers of damage progression and proteinuria development in diabetic nephropathy. Abstract Diabetic nephropathy (DN) is a severe kidney-related complication of type 1 and type 2 diabetes and the most frequent cause of end-stage kidney disease. Extracellular vesicles (EVs) present in the urine mainly derive from the cells of the nephron, thus representing an interesting tool mirroring the kidney’s physiological state. In search of the biomarkers of disease progression, we here assessed a panel of urinary EV miRNAs previously related to DN in type 2 diabetic patients stratified based on proteinuria levels. We found that during DN progression, miR145 and miR126 specifically increased in urinary EVs from diabetic patients together with albuminuria. In vitro, miRNA modulation was assessed in a model of TGF-β1-induced glomerular damage within a three-dimensional perfusion system, as well as in a model of tubular damage induced by albumin and glucose overload. Both renal tubular cells and podocytes undergoing epithelial to mesenchymal transition released EVs containing increased miR145 and miR126 levels. At the same time, miR126 levels were reduced in EVs released by glomerular endothelial cells. This work highlights a modulation of miR126 and miR145 during the progression of kidney damage in diabetes as biomarkers of epithelial to mesenchymal transition. | Upregulation of miR145 and miR126 in EVs from Renal Cells Undergoing EMT and Urine of Diabetic Nephropathy Patients
Diabetic nephropathy is one of the most frequent complications of diabetes, resulting from diffuse damage to different kidney cells. The identification of subjects at risk is mandatory to prevent its development and provide appropriate therapies reducing the unmanageable evolution towards end-stage kidney disease. The aim of this work was to identify urinary-derived extracellular vesicles (EVs) miRNA cargo to be used as biomarker of kidney damage in diabetic patients. The miRNA profile was then correlated with the molecular mechanism associated with the glomerular and tubular damage using a diabetic-like model. In patients, miR145 and miR126 in urinary EVs increased together with albuminuria. MiR145 and miR126 increased in parallel in EVs from renal epithelial cells undergoing transition to a fibrotic mesenchymal phenotype. These data unveiled a role for miR126 and miR145 as the biomarkers of damage progression and proteinuria development in diabetic nephropathy.
Diabetic nephropathy (DN) is a severe kidney-related complication of type 1 and type 2 diabetes and the most frequent cause of end-stage kidney disease. Extracellular vesicles (EVs) present in the urine mainly derive from the cells of the nephron, thus representing an interesting tool mirroring the kidney’s physiological state. In search of the biomarkers of disease progression, we here assessed a panel of urinary EV miRNAs previously related to DN in type 2 diabetic patients stratified based on proteinuria levels. We found that during DN progression, miR145 and miR126 specifically increased in urinary EVs from diabetic patients together with albuminuria. In vitro, miRNA modulation was assessed in a model of TGF-β1-induced glomerular damage within a three-dimensional perfusion system, as well as in a model of tubular damage induced by albumin and glucose overload. Both renal tubular cells and podocytes undergoing epithelial to mesenchymal transition released EVs containing increased miR145 and miR126 levels. At the same time, miR126 levels were reduced in EVs released by glomerular endothelial cells. This work highlights a modulation of miR126 and miR145 during the progression of kidney damage in diabetes as biomarkers of epithelial to mesenchymal transition.
Diabetic nephropathy (DN) is one of the most frequent and severe chronic complications of diabetes, characterized by persistent high albuminuria and a subsequent decline in the glomerular filtration rate [1,2]. The changes in kidney function are associated with specific histopathological findings in glomerular and tubulointerstitial compartments, with renal cell hyperplasia and hypertrophy, the thickening of glomerular and tubular basement membranes and the expansion of tubulointerstitial and mesangial compartments [3,4]. It is well established that patients suffering from both type 1 and type 2 diabetes develop nephropathic complications very early in the progression of the disease; thus, the identification of subjects at risk of DN is required to provide appropriate therapy and slow down evolution towards end-stage renal disease [5]. Currently, the assessment of DN is based on the levels of retention markers, such as creatinine and urea, and on the amount of proteins found in urine. Indeed, albuminuria is today one of the most important indexes applied to assess the progression of DN and one of the main prognostic factors for renal interstitial fibrosis [6]. On the other side, the degree of proteinuria does not accurately reflect the severity or the prognosis of DN, nor is it specific, being a hallmark of glomerular dysfunction [6]. In addition, tissue damage and the induction of inflammation have already occurred by the time that albuminuria is detectable. Finally, a decline in kidney function in diabetic patients can occur in the presence of non-albuminuric or non-proteinuric DN [7], making it difficult to establish the correct timing to initiate the appropriate therapeutic intervention. This intensifies the need for sensitive non-invasive biomarkers of DN [5]. Extracellular vesicles (EVs) are small particles composed of a lipid bilayer secreted by all cell types under both physiological and pathological conditions. EVs display a relevant role in cell signaling and communication, thanks to the potential to transfer their molecular cargo (proteins, lipids and nucleic acids) and play a central function in kidney physiology and pathology [8]. In particular, EVs present in urine (uEVs) carry molecules that are characteristic of the epithelial cells present in the whole length of the urinary tract [9]. Accordingly, uEVs are attracting increasing interest as potential urinary biomarkers and may represent a valuable source of information related to the state of renal tissue. Indeed, several data have highlighted the possible use of uEV cargo, including miRNAs, as biomarkers of kidney diseases [8,9], possibly reflecting their dysregulation in the renal tissue at various stages of DN [10]. In fact, miRNAs are relatively stable in tissue and biological fluids, particularly when carried by EVs [11]. However, the diagnostic application of those biomarkers still needs multicenter validation and large patient cohorts [9]. In search of markers of DN progression in preliminary experiments, we here assessed a panel of uEV miRNAs previously related to endothelial injury, tubular damage and renal fibrosis [12,13,14,15,16,17,18,19,20], comparing miRNA levels in diabetic patients stratified based on proteinuria levels. We therefore focused our attention on two miRNAs, miR126 and miR145, that showed an altered expression in patients with DN during disease progression. Moreover, we focused on the identification of the renal cells potentially responsible for the uEV-associated miR release and the biological process behind this phenomenon. In in vitro diabetes-mimicking models generated by hyperglycemic and fibrotic stimulation, miRNA levels released by tubular and glomerular cell-derived EVs correlated with the induction of epithelial to mesenchymal transition (EMT).
EVs were isolated from the urine of 46 diabetic patients following the protocol in [21]. uEVs expressed the classical exosomal markers, tetraspanins (CD63, CD9 and CD81), as evaluated by super resolution microscopy. DN uEVs appeared heterogeneous, expressing single or multiple tetraspanins at single uEV level (Figure 1A). Moreover, the renal origin of uEVs was confirmed by the presence of AQP1 and AQP2, markers of different nephron segments (Figure 1A). TEM analysis reveals the typical size distribution of EVs (Figure 1B) and the typical cap-shaped morphology (Figure 1B inset). Cytofluorimetric analysis confirmed the presence of tetraspanins, as well of classical urinary markers, such as CD24 and CD133 [21] and the epithelial marker CD326. No expression of the endothelial markers CD31 and CD146 was detected (Figure 1C). uEV were then divided based on the urinary albumin level into the following groups: diabetes patients with normoalbuminuria (NAlb DN), patients with microalbuminuria (MiAlb DN) and patients with macroalbuminuria (MaAlb DN). Table 1 summarizes the clinical features of diabetic patients enrolled in the study. In patients with diabetes, only serum creatinine values appeared to be normally distributed, and the t-test showed significant differences among groups (p = 0.04). A group-to-group comparison using the Mann–Whitney test for the other non-normally distributed data evidenced a significant difference in glomerular filtration rate and albuminuria but not in glycated hemoglobin values. No significant difference was observed in uEV size between groups (Table 1).
In search of markers related to DN progression, uEVs from diabetic patients with normo, micro, or macroalbuminuria were analyzed for the expression of a panels of miRNAs previously reported to be involved in renal cell damage or fibrosis (Supplementary Table S1) [12,13,14,15,16,17,18,19,20]. In preliminary experiments, no significant modulation of miR21, miR24, miR221, miR296, and miR320c was observed (Supplementary Table S2). On the contrary, the level of miR145 and miR126 in uEVs from diabetic patients appeared significantly modulated within the groups. In particular, a significant increase of miR126 was observed in proteinuric diabetic patients compared to diabetic patients without kidney disease. No differences were detected between micro- and macroalbuminuric patients (Figure 2A). It is of interest that miR126 was previously reported to be elevated in urine of diabetic patients [20]. In parallel, miR145 levels were significantly higher in uEVs from DN patients in respect to normoalbuminuric patients, and further increased with the development of proteinuria (Figure 2B). This result corroborates and extends the observation that uEV miR145 level positively correlates with the onset of microalbuminuria during diabetes [19].
To assess the possible origin of the increased uEV miRNAs, podocytes and glomerular endothelial cells were subjected to TGF-β1 treatment in a millifluidic glomerular system, to mimic complication associated with long-term diabetes, and EVs were isolated separately from the two compartment supernatants (Figure 3A). Under TGF-β1 stimulation, podocytes underwent EMT with a significant increase in α-SMA, N-cadherin, SLUG and TWIST (Figure 3B) with respect to unstimulated podocytes. In the millifluid system, podocytes after TGF-β1 stimulation showed an increased expression of miR145, while miR126 showed no significant changes (Supplementary Figure S1A). Interestingly, a significant increase in both miR145 and miR126 was detected in EVs derived from podocytes subjected to TGF-β1 (Figure 3C). On the contrary, in EVs derived from glomerular endothelial cells, TGF-β1 stimulation was accompanied by a relevant reduction in EV-derived miR126 levels (Figure 3D), whereas miR145 was absent in all the conditions tested (not shown). No significant changes in miR126 or miR145 levels were observed in endothelial cells when treated with TGF-β1 (Supplementary Figure S1B).
As proximal tubular cells represent the other potential source of EVs modulated in urine in diabetic conditions, we set up a model of diabetic-induced tubular damage. The tubular cell line HK2 was subjected to glucose and albumin (HSA) overload, to mimic diabetic and proteinuric conditions. As observed in podocytes, renal tubular cells under this condition underwent EMT, showing a significant decrease in the E-cadherin and an increase in α-SMA and TWIST (Figure 4A). No significant modulation of miR126 and miR145 was observed in renal tubular cells subjected to glucose and albumin overload (Figure 4B). At variance, in the EVs released by HK2 submitted to diabetic-induced damage, a significant enhancement of miR145 and miR126 was detected (Figure 4C).
EVs are nanosized particles constantly secreted by all cells and their features reflect the state of the cell of origin, so that they can mirror tissue health and disease [22,23,24]. From a clinical perspective, EVs can be used as appropriate biomarkers to evaluate disease evolution. In the present study, we show that the uEV levels of miR126 and miR145 change in diabetic conditions during the progression of the renal damage. Moreover, we confirmed the specific increase of those two miRNAs in EVs released by podocytes and not glomerular endothelial cells in diabetic-like conditions under millifluidic perfusion, taking advantage of a 3D glomerular model, as well as by tubular epithelial cells under glucose and protein overload. Recently, alterations in miRNA expression profiles have been associated with several pathological processes supporting an increasing interest in their exploitation in the diagnosis and prognosis of pathological conditions [11,25]. Here, analyzing several EV associated miRNAs known to be altered in diabetes and DN, we specifically identified two miRNAs, miR126 and miR145, modulated in uEVs during disease progression. MiR126 is known to play an important role in maintaining endothelial cells and angiogenesis, as it is a key regulator of endothelial inflammation and maintains vascular homeostasis [26,27]. In diabetes, miR126 levels were reported to have opposite expression in EVs from urine in respect to those from serum. Indeed, recent findings showed that miR126 levels in uEVs were significantly enhanced in diabetic patients with kidney disease compared to ones without kidney damage [20]. In parallel, miR126 levels were reported as significantly reduced in serum EVs in diabetic conditions [17]. In the present study, we found that miR126 levels were significantly different between uEVs from non albuminuric and uEVs from DN patients with both micro and macroalbuminuria. These results indicate that this marker might be specifically related to the progression of diabetic condition and modulated during the kidney damage. The data observed in uEVs from diabetic patients paralleled the results obtained in our in vitro models of diabetes-induced damage. We found that EVs released by podocytes and tubular cells undergoing EMT showed an increase in miR126 levels. In parallel, in our 3D model, glomerular endothelial cells co-cultured in the milli-fluidic glomerular system and submitted to TGF-β1 stimulation showed a marked decrease in miR126 EV level, with no significant changes observed in cells. These data suggest that miR126 packaging in endothelial and renal epithelial cell-derived EVs released in serum can be positively or negatively modulated by pathological stimuli, such as TGF-β1, as well as albumin or glucose overload [17,28]. Our data linking EMT with uEV miR126 levels in renal epithelial cells are supported by previous studies showing that the upregulation of miR126 may promote the development of liver fibrosis in hepatic stellate cells [29]. Moreover, in systemic sclerosis, miR126 has been demonstrated to contribute to the downregulation of the angiogenic factor EGFL7 and may influence fibrosis via collagen modulation [30]. We also found that miR145 levels increased in uEVs from patients with DN with respect to patients with diabetes without renal complication. This result validates in a different cohort the previous observation that in uEVs miR145 positively correlated with the onset of microalbuminuria during diabetes [19]. We here extended this finding in macroalbuminuric subjects in which miR145 was further increased. This result highlights a link for miR145 and disease progression. However, the role of this miRNA in renal tissue damage is unclear, as miR145 can positively or negatively regulate fibrosis in different pathological processes [31,32,33,34,35]. In tumor cells, miR145 prevented cell invasion and EMT [32,33,34]. At variance, in peritoneal fibroblasts, TGF-β1-induced miR145 accounted for EMT induction and fibrosis development through FGF10 decrease [31]. Similarly, in podocytes, miR145 enhancement was recently correlated with foot process effacement and the development of proteinuria [35]. These effects were due to Rho-related pathway targeting and subsequent increase in Rac1 and Cdc42 activity, followed by podocyte injury [35]. In our in vitro experiments, the TGF-β1 treatment of podocytes-induced miR145 expression and determined the release of EVs enriched of this miRNA. In addition, an increase of miR145 was also present in EVs from tubular cells challenged with glucose and albumin overload, in diabetic-like conditions. Beside the role as biomarkers, the possible functional effect of EV-contained miR126 and miR145 would be of interest, considering that both miRNAs may affect EMT positively and negatively [26,27,28,29,30,31,32,33,34,35,36]. It could be speculated that uEV cargo during DN may be involved in the amplification of diabetic-induced alterations via EV release or, alternatively, simply represent the modulation of the miRNA packaging in EVs from damaged renal cells. Understanding these intracellular mechanisms and precisely following the axis of miRNA-messenger RNA in kidney cells [23,24,25,36] is crucial for the future use of EVs in the clinical evaluation of DN. Further experiments are needed to clarify this aspect. Altogether, our data indicate that miR126 and miR145 are regulated within the released EVs during the development of the fibrotic damage during diabetes occurrence, supporting their possible role in further patients’ stratification. In addition, we link the presence of those miRNAs within uEVs with their release by podocytes and tubular epithelial cells under diabetes-mimicking conditions.
All patients enrolled in the present study provided informed written consent for the study. The study protocol was approved by the Bioethics Committee of the A.O.U. Città della Salute e della Scienza Hospital (protocol no. 0021671). The study was conducted according to the principles expressed by the Declaration of Helsinki of 1975, as revised in 2013. The study group was composed of a total of 46 adult patients with type 2 diabetes admitted to the clinic (HbA1c > 48 mmol/mol) and divided based on their albuminuria levels in normoalbuminuria, microalbuminuria and macroalbuminuria patients.
Morning urine samples (~100 mL) were collected in sterile containers. In parallel, biochemical analysis was performed by the clinical laboratory of the A.O.U. Città della Salute e della Scienza Hospital. Urine samples were centrifuged at 3000 rpm for 15 min to remove whole cells, large membrane fragments and other debris. Protease Inhibitor (PI) Cocktail (Sigma-Aldrich. St. Louis, MO, USA, 100 μL PI/100 mL urine) and NaN3 (10 mM, Sigma-Aldrich) were added immediately to the remaining supernatant. After filtration through 0.8- and 0.45-µm filters (Merck Millipore, Burlington, MA). uEVs were collected from the samples through ultracentrifugation (Beckman Coulter, OPTIMA L-100 K Ultracentrifuge, Rotor Type 70-Ti, Brea, CA, USA) at 100,000× g for 1 h at 4 °C, as described [21]. The pellet was then resuspended in RPMI (Euroclone, Turin, Italy) + 1% DMSO (Sigma-Aldrich) and stored at −80 °C until use.
HK2 cells (ATCC) were cultured in Dulbecco’s Modified Eagle Medium Low Glucose (DMEM LG) supplemented with 10% foetal bovine serum (FBS) (Euroclone S.p.A., Pero, MI, Italy). Immortalized human podocytes, previously characterized [22] were cultured in DMEM High Glucose (Euroclone) with 10% FBS. Glomerular endothelial cells (GEC) (Cell Biologics Inc., Chicago, IL, USA) previously immortalized [22] were cultured in EndoGRO-LS Complete Culture Media (Merck Millipore) and 10% FBS. Penicillin-Streptomycin (PS) and L-glutamine were added to the cell cultures.
For the glomerular model of diabetic induction, a millifluidic device (IVtech Srl., Lucca, Italy) was used as previously described [22]. The model was assembled in a structure in which GEC and podocytes were separated with a layer of collagen IV (Sigma-Aldrich). Diabetic induction was achieved by adding Transforming Growth Factor-β1 (TGF-β1) (Sigma-Aldrich, 30 ng/mL) in EV-deprived complete culture medium for 24 h. For the tubular model, HK2 were treated with human serum albumin (HAS, 10 mg/mL) and high glucose (HG, 30 mM) for 72 h in RPMI 1640 (Gibco BRL, Paisley, UK). Control cells remained in basal DMEM LG.
EVs were collected from the supernatant of kidney cells after 24 h treatment (in the presence of complete medium deprived of EVs by ultracentrifugation). Following the removal of cell debris and apoptotic bodies by centrifugation at 3000× g for 15 min and microfiltration over a 0.22-μm filter. Medium was then centrifuged at 4000 rpm for 15 min with the 3 KDa Amicon filters (Amicon-Merck Millipore). EVs were then purified using the ExoQuick (SystemBio, Palo Alto, CA, USA) following the manufacturer’s protocol. Briefly, ExoQuick was added to the EV containing supernatant and incubated for 2 h at 4 °C. The mixture was then centrifuged at 1500× g for 30 min to separate supernatant and the EV-enriched pellet.
Transmission electron microscopy (TEM) was performed on EVs placed on 200-mesh nickel formvar carbon-coated grids (Electron Microscopy Science) for 20 min to promote adhesion. The grids were then incubated with 2.5% glutaraldehyde plus 2% sucrose. EVs were negatively stained with NanoVan (Nanoprobes, Yaphank, NY, USA) and observed using a Jeol JEM 1400 Flash electron microscope (Jeol, Tokyo, Japan) [37].
EVs were analyzed using nanoparticle tracking analysis using the NanoSight NS300 system (NanoSight, Salisbury, UK) configured with a blue 488-nm laser and a high-sensitivity digital camera system (OrcaFlash 2.8, Hamamatsu C1 1440, NanoSight). Briefly, EVs stored in −80 °C were thawed, strongly vortexed and properly diluted in saline solution (Fresenius Kabi, Bad Homburg, Germany) previously filtered with a 0.1-µm filter (Merck Millipore). For each sample, three videos of 30 s each were recorded. The settings of acquisition and analysis were optimized and kept constant between samples.
A super-resolution experiment was performed using a Nanoimager S Mark II microscope from ONI (Oxford Nanoimaging, Oxford, UK) equipped with a 100×, 1.4 NA oil immersion objective, an XYZ closed-loop piezo 736 stage and triple emission channels split at 488, 555 and 640 nm. For the evaluation of tetraspanin expression, the EV MAN profiler Kit (ONI) was used following the manufacturer’s protocol. Fluorescent antibodies anti CD9-488, CD63-568 and CD81-647 were included in the kit. For the evaluation of renal markers, 2.5 µg of purified rabbit anti-Aquaporin (AQP) 1 and AQP2 (Santa Cruz) were conjugated with Alexa Fluor 647 dye, using the Apex Antibody Labelling Kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. AQP1 and AQP2 were used in the EV MAN profiler Kit in a combination of CD63-568 present in the kit. The samples were washed twice with PBS and 10 μL of ONI Bcubed Imaging Buffer was added for acquisition. Three or two-channel dSTORM data were acquired sequentially at 30 Hz (Hertz) in total reflection fluorescence (TIRF) mode. Single molecule data was filtered using NimOS (Version (v.1.18.3, ONI) based on the point spread function shape, photon count and localization precision to minimize background noise and remove low-precision localizations [38].
EV surface markers were evaluated by cytofluorimetric analysis using the MACSplex Exosome Kit (Miltenyi). A total number of 1 × 109 EVs were used following the manufacturer’s protocol [24]. The acquisition was performed with a BD FACS Celesta Cell analyzer (BD, Franklin Lakes, NJ, USA). The evaluation of the exosome marker levels was calculated first by subtracting the background fluorescence and then normalizing the median fluorescence of each marker with the mean of the median fluorescence of the tetraspanin markers (CD63-CD81-CD9). Acquisition and analysis settings were optimized and kept constant between samples.
Total RNA, both from patients’ urine- and cell-derived EVs, was extracted using the mirVana kit (Thermo Fisher Scientific. Waltham, MA, USA) as per the manufacturer’s protocol. Isolated RNA was quantified using the NanoDrop2000 spectrophotometer (Thermo Fisher Scientific) and either used immediately or stored at −80 °C until further use.
For the gene expression analysis in kidney cells, first-strand cDNA was produced from 200 ng of total RNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). Quantitative Real-time (qRT)-PCR experiments were performed in the 20-μL reaction mixture containing 5 ng of cDNA template, the sequence-specific oligonucleotide primers (purchased from MWG-Biotech. Table 2) and the Power SYBR Green PCR Master Mix (Thermo Fisher Scientific). GAPDH mRNA was used to normalize the RNA inputs. EV miRNA analysis was carried using the miRCURY LNA™ Universal RT microRNA PCR kit (Qiagen, Düsseldorf, Germany). In selected experiments, the spike-in UNISP6 was added during the RNA isolation as additional extraction control. Fifty pg of reverse transcription reaction products were combined with SYBR Green Master Mix (Qiagen) and LNA™ PCR primer mix and analyzed as described by the manufacturer’s protocol. RNU6B or UNISP6 were used as qRT-PCR loading controls based on their stability in the different conditions tested.
Statistical analyses were performed using Graph Pad Prism version 5.04 (Graph Pad Software Inc., La Jolla, CA, USA). Comparison between groups were either analyzed using non-parametric (Mann–Whitney) or parametric (Student’s t-test and one-way ANOVA with Dunnett’s multiple comparison test), when appropriate. A p value of <0.05 was considered significant.
In conclusion, the changes of miRNA expression in uEVs can be a useful tool for the diagnosis of the different stages of DN. The present study first validates the increase in miR126 and miR145 in uEVs in DN and, more importantly, correlates their increase with the progression of DN, as assessed by proteinuria levels. Moreover, we here were confirmed their altered release in EVs derived from renal cells undergoing diabetes-mimicking damage. It can be speculated that the simultaneous evaluation of the two biomarkers may increase the sensitivity of detecting the renal damage progression in this pathology. | true | true | true |
|
PMC9603323 | Laura Díez-Ricote,Paloma Ruiz-Valderrey,Víctor Micó,Ruth Blanco,Joao Tomé-Carneiro,Alberto Dávalos,José M. Ordovás,Lidia Daimiel | TMAO Upregulates Members of the miR-17/92 Cluster and Impacts Targets Associated with Atherosclerosis | 11-10-2022 | cardiovascular disease,TMAO,microRNAs,atherosclerosis,inflammation | Atherosclerosis is a hallmark of cardiovascular disease, and lifestyle strongly impacts its onset and progression. Nutrients have been shown to regulate the miR-17/92 cluster, with a role in endothelial function and atherosclerosis. Choline, betaine, and L-carnitine, found in animal foods, are metabolized into trimethylamine (TMA) by the gut microbiota. TMA is then oxidized to TMAO, which has been associated with atherosclerosis. Our aim was to investigate whether TMAO modulates the expression of the miR-17/92 cluster, along with the impact of this modulation on the expression of target genes related to atherosclerosis and inflammation. We treated HepG-2 cells, THP-1 cells, murine liver organoids, and human peripheral mononuclear cells with 6 µM of TMAO at different timepoints. TMAO increased the expression of all analyzed members of the cluster, except for miR-20a-5p in murine liver organoids and primary human macrophages. Genes and protein levels of SERPINE1 and IL-12A increased. Both have been associated with atherosclerosis and cardiovascular disease (CDVD) and are indirectly modulated by the miR-17-92 cluster. We concluded that TMAO modulates the expression of the miR-17/92 cluster and that such modulation could promote inflammation through IL-12A and blood clotting through SERPINE1 expression, which could ultimately promote atherosclerosis and CVD. | TMAO Upregulates Members of the miR-17/92 Cluster and Impacts Targets Associated with Atherosclerosis
Atherosclerosis is a hallmark of cardiovascular disease, and lifestyle strongly impacts its onset and progression. Nutrients have been shown to regulate the miR-17/92 cluster, with a role in endothelial function and atherosclerosis. Choline, betaine, and L-carnitine, found in animal foods, are metabolized into trimethylamine (TMA) by the gut microbiota. TMA is then oxidized to TMAO, which has been associated with atherosclerosis. Our aim was to investigate whether TMAO modulates the expression of the miR-17/92 cluster, along with the impact of this modulation on the expression of target genes related to atherosclerosis and inflammation. We treated HepG-2 cells, THP-1 cells, murine liver organoids, and human peripheral mononuclear cells with 6 µM of TMAO at different timepoints. TMAO increased the expression of all analyzed members of the cluster, except for miR-20a-5p in murine liver organoids and primary human macrophages. Genes and protein levels of SERPINE1 and IL-12A increased. Both have been associated with atherosclerosis and cardiovascular disease (CDVD) and are indirectly modulated by the miR-17-92 cluster. We concluded that TMAO modulates the expression of the miR-17/92 cluster and that such modulation could promote inflammation through IL-12A and blood clotting through SERPINE1 expression, which could ultimately promote atherosclerosis and CVD.
Nutrition plays a crucial role in the development of non-communicable diseases and overall health. Approximately 45% of deaths caused by cardiometabolic events are associated with poor lifestyle habits and nutrition [1]. Saturated fats and red meat have previously been associated with cardiovascular disease (CVD); however, their specific involvement in its progression has not yet been established [2]. Indeed, some meta-analyses have not found any relationship between intake of red and processed meat and CVDs and all-cause mortality [3]. Choline, betaine (a metabolite from choline), and L-carnitine are abundant in eggs, fish, and meat. They can be metabolized into trimethylamine (TMA) by the gut microbiome. TMA is then absorbed and transported to the liver, where it is metabolized to trimethylamine-n-oxide (TMAO) by the flavin-containing mono-oxygenase (FMO) enzyme [4]. This metabolite has gained attention regarding its link to CVD [5] and has been reported to be a potential CVD marker linked to atherosclerosis and atheroma plaque formation [6]. Plasma TMAO levels could aid in predicting cardiovascular events, even when other biomarkers are within normal ranges [5]. Furthermore, mice supplemented with L-carnitine showed higher TMAO plasma levels and increased atherosclerosis [4], and APOE−/− mice fed with a high-fat diet and TMAO had significantly increased atherosclerotic plaque progression [7]. In addition, TMAO plasma levels are higher in subjects with myocardial infarction and are associated with a higher coronary atherosclerotic load [8]. However, high plasma TMAO levels have also been found in subjects with a healthy diet [9]. In addition, TMAO has not been associated with cardiovascular disease in children and their parents, although its precursors (i.e., choline, carnitine, and betaine) have been [10]. A similar finding was reported in the population from the CARDIA study, where TMAO levels were not associated with measures of atherosclerosis risk in young males and females [11]. Furthermore, TMAO and its relationships with CVDs have been studied in populations at high risk of suffering CVDs, and the debate is still open, so it can be suggested that concentrations of TMAO precursors could be more useful as a biomarker of CVD for the general population. Indeed, TMA has been proposed as a more precise and accurate marker of inflammation and CVD risk, as it has been reported that TMA levels are increased in cardiovascular patients compared to healthy subjects, while no such differences were observed in TMAO [12]. Thus, the TMA/TMAO ratio may be useful when evaluating these metabolites as CVD markers, as might the functionality of the FMO3 enzyme, which is responsible for the oxidation of TMA to TMAO. Very recently, we demonstrated that an increase in the intake of choline or betaine—precursors of TMAO—improved cardiometabolic and renal traits in a population of adult men and women with overweight/obesity and metabolic syndrome in the PREDIMED-Plus study [13]. Another previous study conducted by our team showed that TMAO increases the expression of miR-30c, with a role in lipid metabolism [14,15], and miR-21, which is fundamental in the resolution of acute inflammation [15,16]. In this study, we demonstrated the miR-30c-mediated regulation of PER2—a key circadian regulator—by TMAO and suggested a pro-inflammatory effect of TMAO. Therefore, results are contradictory and, consequently, the effects of TMAO on atherogenesis and cardiometabolic traits need to be further defined. In recent years, more studies have highlighted the role of the miR-17/92 cluster in cardiovascular physiology [17]. Its dysregulation has been associated with arrhythmogenesis and cardiomyopathy [18], aberrant cardiomyocyte differentiation [19], myocardial infarction [20], and cardiac aging [21]. In addition, this cluster is also involved in lipid metabolism [22,23]. Moreover, our previous studies show that members of this cluster are modulated by diet—specifically, by the consumption of extra-virgin olive oil [24] and beer [25]. Given our previous results and the existing literature, we aimed to evaluate whether TMAO could modulate the expression of three microRNAs from the miR-17-92 cluster (i.e., miR-17-5p, miR-20a-5p, and miR-92a-3p) and of relevant targets involved in atherosclerosis progression and CVD, in two cellular models that are important in the biology of inflammation and lipid metabolism (i.e., macrophages and hepatocytes).
To test whether TMAO modified the expression of members of the miR17-92 cluster, we treated cells with 6 µM of TMAO at different times of incubation and analyzed the expression of miR-17-5p, miR-20a-5p, and miR-92a-3p by RT-qPCR. TMAO upregulated miR-17-5p (0.61 ± 0.10 at 8 h and 1.63 ± 0.34 at 24 h in HEPG-2 cells, and 0.79 ± 0.21 at 24 h in THP-1 cells) and miR-92a-3p (0.64 ± 0.12 at 8 h in HEPG-2 cells and 0.34 ± 0.17 at 24 h in THP-1 cells) (Figure 1a,b). These microRNAs were also upregulated in primary human macrophages (miR-17-5p 2.62 ± 0.15; miR-92a-3p 2.59 ± 0.16) and murine liver organoids (miR-17-5p 1.71 ± 0.19 at 8 h and 0.82 ± 0.08 at 24 h; miR-92a-3p 0.32 ± 0.14 at 8 h and 0.27 ± 0.19 at 24 h) (Figure 1c,d). However, miR-20a-5p was only upregulated in HEPG-2 (0.85 ± 0.76, 0.72 ± 0.38, and 1.58 ± 0.27 at 4, 8, and 24 h, respectively) and THP-1 cells (0.65 ± 0.10 at 24 h), but not in human primary macrophages or murine liver organoids (Figure 1c,d). We then searched the literature for targets of these microRNAs that were involved in regulating inflammation, endothelial or cardiac function, and the formation and progression of atheroma plaques. Among the putative targets, we selected SERPINE1 and IL-12 for further study due to their key role in the progression of atherosclerosis and the resolution of inflammation, respectively [26,27,28]. The miR-17-92 cluster indirectly targets SERPINE1 and inhibits its expression by modifying the PDLIM5 gene regulation involved in translocation of kinase to the muscle and cardiomyocyte contraction [29]. In addition, SERPINE1 could inhibit smooth muscle cell proliferation from arteries, stimulated by the miR-17/92 cluster. In neuroblastoma cells, it has been observed that when expression of the miR-17/92 cluster was activated, both TGF-β and SERPINE1 were inhibited [30]. We found that SERPINE1 gene expression was significantly increased and then decreased after 4 h (0.67 ± 0.14) and 8 h (−0.57 ± 0.14) of treatment, respectively, in HEPG-2 cells (Figure 2a). On the other hand, SERPINE1 protein levels were increased at both 8 h (1.73 ± 0.35) and 24 h (2.71 ± 0.9). We observed a significant increase in IL-12A gene expression levels in THP-1 cells incubated with 6 µM of TMAO for 12 h (0.50 ± 0.17), whereas protein levels were significantly increased at the 12 h (1.11 ± 0.16) and 24 h (1.70 ± 0.27) timepoints (Figure 2b).
This study shows how TMAO, at a physiological dose (6 µM), modulates the expression of the miR-17/92 cluster in two cell lines (HEPG-2 and THP-1). Additionally, we validated our results using two non-immortalized cellular models whose characteristics are likely to closely resemble what occurs in vivo, i.e., murine liver organoids and human primary macrophages. We found that miR-17-5p and miR-92a-3p were upregulated after TMAO exposure in HEPG-2 cells and liver organoids at 8 h and 24 h, and in THP-1 cells and human macrophages at 24 h. However, miR-20a-5p did not change in the murine liver organoids or human primary macrophages. The miR-17-92 cluster, also known as oncomiR1 because of its role in the development of cancer, modulates energy metabolism in cancer cells. It has been shown that the lack of miR-92a upregulates glycolytic and oxidative metabolism in cancer cells; however, the lack of miR-17 and miR-20a inhibits glycolytic and oxidative metabolism as well as mTOR pathways in cancer cells and increases the AMPK signaling pathway [31]. Indeed, in silico analyses have shown that these nutrient-sensing pathways are enriched in those processes regulated by miR-17 and miR-20a [24]. This cluster has also been linked to lipid metabolism, being differentially expressed in coronary artery disease (CAD) subjects, as miR-17 and miR-20a were upregulated and other members of the cluster were downregulated [22]. It has also been positively associated with hyperlipidemia and total cholesterol levels [22]. A recent study found that circulating levels of miR-92a were increased in patients with hypertension and were correlated with atherosclerosis markers [32]. Another study found that miR-92a was upregulated in plasma exosomes from subjects with CAD and was suggested to be a potential biomarker for atherosclerosis diagnosis [33]. miR-17 has been found to significantly increase atherosclerosis in patients [34]. miR-17 is also increased in human acute myocardial infarction subjects [35] and mice [36]. Thus, it could be suggested that TMAO may increase cardiovascular risk by increasing the expression of these microRNAs, which could be used as potential biomarkers of cardiovascular disease. Interestingly, we found that TMAO significantly upregulated miR-20a-5p in the two cancer cell lines used in our experiments (THP-1 and HepG-2), while it did not change in primary macrophages or murine liver organoids, suggesting that TMAO could have a different impact depending on the health status and the presence of cancerous processes. The role of miR-20a in cardiovascular diseases is controversial; miR-20a plasma levels are significantly increased in patients who have suffered postinfarction heart failure, and have also been associated with left ventricular dilatation, suggesting that miR-20a could be a potential biomarker for left ventricular modeling in postinfarction patients [37]. Furthermore, ApoE−/− mice with miR-20a knockout exhibited less atherosclerotic formation. In addition, miR-20a promoted atherosclerotic development and reduced reverse cholesterol transport in these mice [38], suggesting that miR-20a inhibition could be a potential target in atherosclerosis therapy. On the other hand, an in vitro study showed that miR-20a was repressed when human aortic endothelial cells were treated with oxidized LDL particles, and overexpressed miR-20a protected cells from ROS generation, suggesting that miR-20a could protect aortic cells from inflammatory conditions and atherosclerotic development [39]. Furthermore, miR-20a is protective against myocardial ischemia/reperfusion injury, as miR-20a mimics promoted cardiomyocyte viability and decreased apoptosis [40]. Thus, as miR-20a behaves differently in cancer cell lines and primary cells, it could be suggested that TMAO may impair lipid homeostasis through miR-20a only in tumor cells. Further research would be needed to clarify the role of miR-20a in cardiovascular disease and atherosclerosis. We only studied SERPINE1 in HEPG-2 cells, as it is not highly expressed in macrophages. SERPINE1 (also known as PAI-1: plasminogen activator inhibitor 1) is involved in normal blood clotting, as it acts as an inhibitor of tissue-type plasminogen activator and urokinase-type plasminogen activator. SERPINE1 is highly expressed in subjects with atherosclerotic plaques [26,27,41], and it is thought to promote atherogenesis due to its pro-thrombotic capacity and its ability to deposit fibrin in atherosclerotic lesions [26]. It has also been reported that SERPINE1 expression correlates with the thickness and severity of atherosclerosis in human arteries [26], and it is localized in endothelial cells, smooth muscle cells, and macrophages [27]. Indeed, SERPINE1 is localized in endothelial and smooth muscle cells in healthy atherosclerotic human arteries. In advanced atherosclerotic lesions, its expression not only increases but is also localized in macrophages [41]. Overexpression of SERPINE1 leads to a reduced remodeling capacity of smooth muscle cells because it inhibits plasmin production, activating metalloproteases needed for remodeling. Furthermore, different cytokines stimulate SERPINE1 production [42]. It has been reported to be linked to hepatic steatosis [43]. SERPINE1 plasma levels are elevated in non-alcoholic fatty liver disease (NAFLD) and metabolic syndrome, positively correlating with VLDL plasma levels, body mass index, and T2D [44]. Furthermore, different SERPINE1 polymorphisms have been associated with a higher risk of CVD [45]. In subjects with thrombotic disorders, SERPINE1 was regulated by plasma miR-30c, which directly inhibited it [46]. Our results suggest that TMAO exposure in the liver could increase cardiovascular risk by increasing SERPINE1 expression, both through its target microRNAs and independently of them, and SERPINE1 could constitute a potential predictor of metabolic-related diseases. IL-12A belongs to the IL-6/IL-12 cytokine family, forming the IL-35 heterodimeric cytokine that plays a role in inflammatory diseases [47]. The IL-12 cytokine family is pro-inflammatory and essential for the progression of hypertension. These cytokines promote interferon-gamma secretion and Th immune response, which are essential for macrophage differentiation [28]. IL-12A polymorphisms have been associated with the risk of coronary artery disease, and it has been suggested that different components of IL-35 (such as IL-12A) could influence the progression of atherosclerosis [47]. IL-12A has also been found to be upregulated in myocardial infarction [48]. TMAO increased IL-12A expression in THP-1 cells at the mRNA and protein levels, especially in the first hours of exposure, suggesting that TMAO promotes inflammation and could promote macrophage differentiation and atherosclerotic plaque progression through IL-12A. Since we found an upregulation of the microRNAs and their genes and proteins in response to TMAO, we may suggest that the effects of TMAO on SERPINE1 and IL-12A are not directly mediated by these microRNAs. SERPINE1 has been suggested to be directly targeted by some members of the miR-17-92 cluster—specifically, by miR-19a/b. However, it has also been suggested that miR-17 and miR-20a could indirectly inhibit SERPINE1 through PDLIM5 [29,30]. Similarly, it has been reported that miR17-92 suppresses IL-12 production in macrophages in genetically modified mice. The modulation of IL-12 by this cluster is mediated by PTEN and, consequently, by the PI3K-Akt-GSK3 pathway [49]. However, higher levels of plasma TMAO were correlated with higher levels of plasma pro-inflammatory markers in peritonitis, including SERPINE1 [50]. Similarly, plasma TMAO levels were positively correlated with plasma IL-12 levels in patients with common variable immunodeficiency [51]. Therefore, TMAO could increase these inflammatory markers independently of the miR-17-92 cluster. The main limitation of our study is its descriptive nature. Further studies are needed to reveal the molecular mechanisms by which TMAO modulates the expression of the miR-17-92 cluster. Furthermore, we did not determine whether the TMAO-mediated regulation of SERPINE1 and IL-12 is directly or indirectly mediated by the miR-17-92 cluster. Although these constraints prevent us from concluding that TMAO modulates SERPINE1 and IL-12 through the miR-17-92 cluster, our results strongly suggest that the miR-17-92 cluster could play a role in the effects of TMAO on inflammation and cardiovascular health.
HepG-2 and THP-1 cells were obtained from the American Type Tissue Collection (Barcelona, Spain). The cells were maintained in RPMI for THP-1 and in DMEM for HepG-2 (Lonza), with 10% fetal bovine serum (FBS) supplemented with glutamine and antibiotics (Cultek). THP-1 cells were differentiated into macrophages by incubating them with 50 ng/mL of phorbol 12-myristate-13-acetate (PMA) for 72 h (Sigma). Peripheral blood mononuclear cells (PBMCs) were obtained from a young, healthy male donor. The study protocol was approved by the IMDEA Food Ethics Committee (PI-040, 5 March 2020, Madrid, Spain), and the donor provided signed informed consent. PBMCs were isolated with Lymphoprep™ (StemCell Technologies) following the manufacturer’s instructions and differentiated into macrophages as previously described in [15]. Murine livers were obtained and their organoids were differentiated as previously described in [15]. All procedures were carried out following the European Communities Directive 86/609/EEC guidelines. The protocols were approved by the Animal Ethics Committee (Proex 281/15 and Proex 282/15) of Ramón y Cajal Hospital (Madrid, Spain).
HEPG-2 cells were treated with 6 µM of TMAO for 4, 8, and 24 h. THP-1 cells were treated with 6 µM of TMAO for 12 and 24 h. Murine liver organoids were treated with 6 µM of TMAO for 4, 8, and 24 h, while human primary macrophages were treated with 6 µM of TMAO for 12 h. Treatment doses and times were selected from dose– and time–response curves previously produced in [15].
Cells were lysed with Tripure (ROCHE, Madrid, Spain). RNA was isolated following the phenol:chloroform methodology adapted to enhance microRNA precipitation [15]. RNA concentration was measured using a NanoDrop 2000 (Thermo Scientific, Rockford, IL, USA), and RNA integrity was checked in 2% agarose gels. Micro-RNA-enriched RNA was isolated from human primary macrophages using the miRCURY RNA Isolation Kit—Cell and Plant (Exiqon, Bionova, Madrid, Spain) following the manufacturer’s instructions. miRNA expression levels were measured by real-time quantitative PCR (RT-qPCR). First, they were reverse-transcribed with the miScript Reverse Transcription II Kit (Qiagen, Madrid, Spain) and amplified with the miScript SYBR Green PCR Kit (Qiagen, Madrid, Spain), using a specific forward primer (Isogen Lifescience, Barcelona, Spain) and a universal reverse primer included in the kits. microRNA levels were calculated with the 2−ΔΔCt method, using RNU6 and RNU43 for normalization.
Gene expression levels were measured by real-time quantitative PCR (RT-qPCR). RNA was reverse-transcribed with the Prime Script Reverse Transcription Kit (Takara, CA, USA), and cDNA was amplified using the FastStart Universal SYBR Green Master (Roche, Barcelona, Spain). Specific forward and reverse primers were used (Isogen Lifesciences, Barcelona, Spain). Gene expression levels were calculated using the 2−ΔΔCt method and normalized with RN18S and RPLP0.
Cells were lysed in ice-cold NP-40 lysis buffer, and total protein was quantified using the Pierce TM BCA Protein Assay Kit (Thermo Scientific, Rockford, IL, USA). Then, 50 µg of protein was separated by SDS–PAGE, transferred onto nitrocellulose membranes, and probed with the anti-IL-12A 1:500 anti-Serpine1 (Biorbyt) 1:2000, anti-β-actin (Abcam, Madrid, Spain) 1:5000 primary antibodies, and the appropriate HRP-conjugated secondary antibodies. Protein bands were visualized and measured via densitometric analyses (ImageJ). Protein quantification was normalized with β-actin.
All experiments were performed in triplicate, and biological triplicates were included in each experiment. Additionally, technical duplicates were included in all RT-qPCR reactions. Inconsistent replicates were repeated. Gene and microRNA expression values were Log2-transformed to adjust the data to a normal distribution. The experiments included a control sample in each timepoint. Therefore, Student’s t-test was used to compare microRNAs, along with gene and protein expression values, between TMAO- and control-incubated samples at each timepoint. Statistical significance was set at p < 0.05. All analyses were performed using SPSS 19.0 (SPSS Inc., IBM Spain, Madrid, Spain).
In summary, we found that miR-17 and miR-92a—two members of the miR-17-92 cluster—were upregulated by TMAO, suggesting that TMAO modulates this cluster. The fact that miR-20a levels were not changed by TMAO means that TMAO could specifically modulate some members of the cluster, but not all, and that miR-20a could respond differently when there are cancerous processes. TMAO also increased the expression of two targets of inflammation and atherosclerosis—SERPINE1 and IL-12A—at the gene and protein levels. However, these inflammatory markers could be modulated by TMAO independently of the miR-17-92 cluster. Thus, it could be suggested that TMAO-mediated induction of the miR-17-92 cluster and of SERPINE1 and IL-12A could promote the development of inflammation and atherosclerosis. | true | true | true |
|
PMC9603387 | Yasunori Fukumoto,Masayoshi Ikeuchi,Yuji Nakayama,Yasumitsu Ogra | Rad17 Translocates to Nucleolus upon UV Irradiation through Nucleolar Localization Signal in the Central Basic Domain | 14-10-2022 | Rad17,DNA damage response,nucleolus,nucleolar localization,subnuclear localization | The nucleolus is a non-membranous structure in the nucleus and forms around ribosomal DNA repeats. It plays a major role in ribosomal biogenesis through the transcription of ribosomal DNA and regulates mRNA translation in response to cellular stress including DNA damage. Rad17 is one of the proteins that initiate and maintain the activation of the ATR pathway, one of the major DNA damage checkpoints. We have recently reported that the central basic domain of Rad17 contains a nuclear localization signal and that the nuclear translocation of Rad17 promotes its proteasomal degradation. Here, we show that the central basic domain contains the nucleolar localization signal as well as the nuclear localization signal. The nucleolar localization signal overlaps with the nuclear localization signal and is capable of transporting an exogenous protein into the nucleolus. Phosphomimetic mutations of the central basic domain inhibit nucleolar accumulation, suggesting that the post-translational modification sites regulate the nucleolar localization. Nucleolar accumulation of Rad17 is promoted by proteasome inhibition and UV irradiation. Our data show the nucleolar localization of Rad17 and suggest a possible role of Rad17 in the nucleolus upon UV irradiation. | Rad17 Translocates to Nucleolus upon UV Irradiation through Nucleolar Localization Signal in the Central Basic Domain
The nucleolus is a non-membranous structure in the nucleus and forms around ribosomal DNA repeats. It plays a major role in ribosomal biogenesis through the transcription of ribosomal DNA and regulates mRNA translation in response to cellular stress including DNA damage. Rad17 is one of the proteins that initiate and maintain the activation of the ATR pathway, one of the major DNA damage checkpoints. We have recently reported that the central basic domain of Rad17 contains a nuclear localization signal and that the nuclear translocation of Rad17 promotes its proteasomal degradation. Here, we show that the central basic domain contains the nucleolar localization signal as well as the nuclear localization signal. The nucleolar localization signal overlaps with the nuclear localization signal and is capable of transporting an exogenous protein into the nucleolus. Phosphomimetic mutations of the central basic domain inhibit nucleolar accumulation, suggesting that the post-translational modification sites regulate the nucleolar localization. Nucleolar accumulation of Rad17 is promoted by proteasome inhibition and UV irradiation. Our data show the nucleolar localization of Rad17 and suggest a possible role of Rad17 in the nucleolus upon UV irradiation.
The nucleolus is a non-membranous structure in the nucleus and forms around ribosomal DNA (rDNA) repeats. It plays a major role in ribosomal biogenesis through the transcription of rDNA by RNA polymerase I and regulates mRNA translation in response to cellular stress including DNA damage [1]. DNA damage response inhibits RNA polymerase I transcription, and the inhibition requires the ATM pathway, one of the major DNA damage checkpoints [2]. Another major DNA damage checkpoint is the ATR pathway, which responds to various chemical forms of DNA damage [3]. The canonical ATR pathway is one of the major checkpoint reactions outside the nucleoli; however, a previous report has disclosed that the ATR pathway is activated in the nucleolus upon inhibition of RNA polymerase I transcription [4], indicating that the ATR pathway also monitors genomic DNA in the nucleolus. Rad17 is one of the proteins that initiate and maintain the activation of the ATR pathway. Rad17 loads the Rad9–Hus1–Rad1 complex (9–1–1 complex) onto damaged chromatin to activate ATR kinase activity. Rad17 also interacts with the Mre11-Rad50-NBS1 complex to activate ATM kinase [5]. The nuclear localization of endogenous Rad17 in non-irradiated [5,6,7] and irradiated [5,8] cells has been reported, and Rad17-S645 phosphorylation signal has been observed in the nucleus [9,10]. By contrast, the nucleolar localization of Rad17 has been poorly characterized. There is one report of the nucleolar staining of endogenous Rad17 [11]; however, the staining was accomplished with only one monoclonal antibody clone, and the possibility of nonspecific staining was not excluded. Other reports have indicated that Rad17 is primarily localized in the nucleoplasm but not in the nucleolus [12,13]. Human Rad17 has a cluster of basic residues, N339–D380, which we named the central basic domain, and this domain is located between N-terminal ATPase and C-terminal α-helical domains (Figure 1A). We have recently reported that the central basic domain contains a nuclear localization signal and that the nuclear translocation of Rad17 promotes its proteasomal degradation. We have also identified tandem destruction boxes of Rad17 on its N-terminus as a set of canonical and noncanonical sequences [14]. The proteasomal degradation of Rad17 is mediated by an anaphase-promoting complex associated with Cdh1 [8], and the N-terminal destruction boxes of Rad17 interact with Cdh1 in vitro [14]. Here, we show that a central basic domain contains the nucleolar localization signal as well as the nuclear localization signal. The nucleolar localization signal overlaps with the nuclear localization signal and is capable of transporting an exogenous protein into the nucleolus. The basic K/R-rich motifs are the key determinants of protein localization in nuclear sub-compartments [15]. On the other hand, the positively charged K/R-rich arrays are nuclear PI(4,5)P2 recognition motifs that are essential for protein localization in nucleoli and nuclear speckles [16]. Indeed, the nucleolar PI(4,5)P2 is in the close proximity to crucial nucleolar constituent fibrillarin [17]. The K/R-rich motif containing proteins were identified in proteins, which are linked proteasomal degradation in an MS-based quantitative approach [16]. Phosphomimetic mutations of the central basic domain inhibit the nucleolar accumulation, suggesting that post-translational modifications regulate the nucleolar localization. Furthermore, UV irradiation promotes the nucleolar accumulation of Rad17, suggesting a nucleolar function of Rad17 in the DNA damage response. Our data show the nucleolar localization and the nucleolar localization signal of Rad17 and suggest a possible role of Rad17 in the nucleolus upon UV irradiation.
In our recent study, we found that EGFP fused with Rad17 E295–D380 peptide showed exclusive nuclear localization [14]. In this study, we further characterized the central basic domain of Rad17 spanning N339–D380 (Figure 1A). Rad17 E295–D380 and E295–E426 peptides were fused with EGFP, and their localization was examined. EGFP-Rad17 E295–D380 was exclusively localized in the nucleus and predominantly accumulated in the nucleolus (Figure 1B, cyan arrowheads). Almost all of the EGFP-positive cells showed the same localization pattern. The nucleolus is made up of three components: the granular component, the dense fibrillar component, and the fibrillar center [18]. In the nucleolus, EGFP-Rad17 E295–D380 surrounds the UBF signal, a marker of the nucleolus fibrillar center. EGFP-Rad17 E295–E426 was also exclusively localized in the nucleus where it was localized solely in the nucleolus (Figure 1B, magenta arrowheads) or distributed in the nucleolus and the nucleoplasm (Figure 1B, yellow arrowheads). The ratios of cells with exclusive nucleolar localization, indicated by the magenta arrowheads, to EGFP-positive cells were 34% and 22% in each experiment. These data indicate that the central basic domain of Rad17 encodes a nucleolar localization signal as well as a nuclear localization signal.
In our previous work, we showed that Rad17 K359–K363 encoded a part of the nuclear localization signal and that K359A/R360A/R361A/K362A/K363A (K/R359–363A or 5KR) mutation abolished the nuclear localization of Rad17 [14]. Here, we examined the effect of K/R359–363A mutation on the nucleolar localization. EGFP-Rad17 E295–D380 having the wild-type sequence (WT) showed 149% accumulation in the nucleolus (referred to as No) relative to the nucleoplasm (Figure 2A,B and Figure S1A). The K/R359–363A mutant of this construct was deficient in the accumulation in the nucleolus and equally distributed in the nucleolus and the nucleoplasm. This mutation also increased the cytoplasmic localization (42%, Figure 2A,B and Figure S1A,B) relative to WT, as was shown recently [14]. EGFP alone was equally distributed in the nucleolus and the nucleoplasm, and no specific localization in the nucleoplasm or the cytoplasm was observed (Figure 2A,B and Figure S1A). The K359A/R360A mutant showed a slight accumulation in the nucleolus (110%) and a significant increase in cytoplasmic localization (33%, Figure 2A,C and Figure S1C,D). The K362A/K363A mutation had a milder effect; it decreased the nucleolar localization (120%) and slightly increased the cytoplasmic localization (13%, Figure 2A,C and Figure S1C,D). These findings indicate that K359 and R360 are central to the nuclear and nucleolar localization signals. We noted that in some cases, a nucleolar localization signal seemed to overlap with a nuclear localization signal [18]; however, we could not differentiate them. Our finding suggests that both signals overlapped in the central basic domain of Rad17. Together, the data indicate that Rad17 K359–K363 residues encode the nucleolar localization signal as well as the nuclear localization signal.
The phosphorylation of Rad17-S348, S351, and S356 residues was confirmed by mass spectrometric analyses and registered in PhosphoSitePlus (https://www.phosphosite.org). In our recent work, we noted that S348D/S351D/S356D mutation decreased the nuclear localization of flag-Rad17 full-length protein [14]. In the EGFP-Rad17 E295–D380 protein, the S348D/S351D/S356D mutation resulted in a decrease in nucleolar accumulation (121%, Figure 2D,E and Figure S1E,F) but did not affect the cytosolic intensity of the EGFP signal (Figure 2D,E and Figure S1F). These results indicate that the phosphorylation sites in the central basic domain regulate the nucleolar localization signal of Rad17.
In our recent work, we found that the nuclear translocation of Rad17 promotes the proteasomal degradation of Rad17 and that the degradation is mediated by N-terminal destruction boxes that interact with Cdh1 [14]. It was also shown that Cdh1 is localized in the nucleus but not in the nucleolus [19]. Here, we examined the relationship between proteasome and the nucleolar localization of Rad17. The Rad17 N-terminal destruction boxes (H36–G66) were fused with EGFP-Rad17 E295–E426 peptide, and the localization was examined. Again, E295–E426 peptide translocated the fused protein to the nucleolus (Figure 3A,B). We examined the effect of mutations in the Rad17 destruction boxes, K36A/P42A/R55A/L58A (KPRL); however, we obtained a marginal result in our preliminary experiments. We also compared protein amount and stability between flag-D box-EGFP-Rad17 E295–D380 and E295–E426; however, we observed small or no difference (data not shown). Then, we examined the effect of proteasome inhibition on the nucleolar localization. Exposure to proteasome inhibitor MG132 promoted the nucleolar accumulation of flag-D box-EGFP-Rad17 E295–E426 (Figure 3A,B). The MG132 exposure also promoted the nucleolar accumulation of flag-EGFP-Rad17 full-length protein (Figure 3C,D). These results suggest that proteasomal degradation negatively regulates the nucleolar localization of Rad17. Our result is consistent with a previous observation that proteasome inhibition induced the nucleolar accumulation of nuclear proteins including ATM [20]. Our current observations suggest that at least two mechanisms regulate the nucleolar localization of Rad17. The first mechanism is the negative regulation by phosphorylation of the central basic domain (Figure 2D,E). The second mechanism is the negative regulation by proteasomal degradation (Figure 3). We previously showed that the nuclear localization of Rad17 is dependent on the nucleotide binding of the Rad17 ATPase domain [21]. Because the nuclear and nucleolar localization signals overlapped in the central basic domain (Figure 2A–C), the Rad17 ATPase domain may also regulate the nucleolar localization signal as the third mechanism.
A previous report has demonstrated that Rad9B, a paralog of canonical Rad9 protein, translocates to the nucleolus upon UV irradiation [13]. Thus, we examined the effect of UV irradiation on the subnuclear localization of Rad17. The flag-EGFP-Rad17 full-length protein accumulated in the nucleolus of a subset of UV-irradiated cells (Figure 4A,B). Rad17 formed discrete foci in the nucleolus or distributed within the nucleolus. Because UV irradiation inhibits general transcription in the nucleus, one possible explanation may be that UV irradiation induces the nucleoplasmic depletion of Rad17 to result in the nucleolar accumulation. However, the inhibition of proteasomal degradation promoted the nucleolar accumulation of Rad17 (Figure 3), suggesting that the degradation or repression of Rad17 does not promote the nucleolar accumulation. These results indicate that UV irradiation induces the nucleolar translocation of Rad17. Rad9B interacts with Hus1, Rad1, and Rad17 but not with TopBP1, suggesting that it is not involved in the activation of ATR and the ATR-dependent DNA damage checkpoint [13]. To our knowledge, Rad9B, Hus1, and Rad1 do not have nucleolar localization signals. In Schizosaccharomyces pombe, Rad17 is required for the nuclear localization of Hus1 and Rad9 [22]. Rad17 may play a role in the translocation of Rad9B to the nucleolus. Upon DNA double-strand breaks, Rad17 directly interacts with NBS1 in the nucleoplasm [5], and NBS1 translocates to the nucleolus to inhibit rDNA transcription [23,24]. Rad17 and Rad9B may be involved in the nucleolar function of NBS1. The physiological function of Rad17 and Rad9B in the nucleolus is still a conundrum; however, our findings suggest the possibility that Rad17 and Rad9B cooperatively play a role upon UV irradiation. Further work will reveal the regulation and the nucleolar function of the Rad17 and Rad9B–Hus1–Rad1 complex.
The following antibodies were used: anti-UBF, F-9, Santa Cruz Biotechnology, sc-13125; rabbit anti-FLAG antibody, Medical & Biological Laboratories, PM020; anti-Hsc70, Santa Cruz Biotechnology, sc-7298; and anti-NPT2, Abcam, ab33595.
The amino acid residues of Rad17 were denoted according to isoform 1 (NCBI NP_579921.1). The pcDNA4 vectors encoding EGFP fused with Rad17 E295–D380 peptide (EGFP-Rad17 E295–D380) were described previously [14]. An EGFP fused with Rad17 E295–E426 peptide (EGFP-Rad17 E295–E426) was constructed in the same manner. The pcDNA3 vectors encoding flag-EGFP Rad17 full-length protein and flag-EGFP were described previously [14]. The pTwist CMV BetaGlobin WPRE Neo vectors encoding flag-D box-EGFP-Rad17 E295–D380 or E295–E426 peptide were synthesized by Twist Biosciences Inc. (San Francisco, CA, USA). The Rad17 H36–G66 sequence that contains tandem destruction boxes (D-box) was inserted between the flag tag and EGFP.
We examined the co-localization of UBF and EGFP fused with the central basic domain of Rad17 as described previously [14,25]. COS-7 cells were transfected with 0.5 μg of pcDNA4/EGFP-Rad17 E295–D380 or E295–E426 peptide using Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA). The cells were fixed with 2% paraformaldehyde 24 h after transfection. The cells were stained with anti-UBF antibody and Alexa Fluor 555 Donkey anti-mouse IgG antibody (Thermo Fisher Scientific) in PBS (−)/3% BSA/0.1% saponin. DNA was stained with 1 μM Hoechst 33342. Fluorescence microscopic images were captured with an IX83 inverted fluorescence microscope (Olympus, Tokyo, Japan). We examined the subcellular and subnuclear localization of EGFP-Rad17 E295–D380, flag-EGFP Rad17 full-length protein, and flag-D box-EGFP-Rad17 E295–E426, as follows. COS-1 cells were transfected with 1.0 μg of plasmids using the acidified polyethylenimine [26]. The cells were fixed with 2% paraformaldehyde 48 h after transfection. To inhibit proteasomal degradation, the cells were exposed to 40 μM MG132 for 7 h before fixation. To examine the effect of UV irradiation, the cells were irradiated with 30 J/m2 of UV-C and allowed to recover for 3 h before fixation. The cells were treated with 200 μg/mL RNase A for 1 h and stained with 5 μg/mL propidium iodide for 30 min. The data were obtained with an LSM 700 or an LSM 5 Pa deconvolution microscope (Carl Zeiss, Jena, Germany). The average intensity of EGFP in the nucleolus, nucleoplasm, and cytoplasm was quantitated with ZEN 3.4 (blue edition) or Image J, and the average intensity ratio was calculated. The intensity of the nucleoplasm was used as 100% standard. The position of the nucleolus was determined on differential inference contrast or phase contrast. The dot and box–whisker plots were written with matplotlib v3.4.3 and seaborn v0.11.2. Whiskers represent the highest and lowest data, excluding outliers, and boxes represent 25%, 50%, and 75% percentiles. Student’s, Welch’s, or one-sample t-test was performed with the stat module of NumPy v.1.21.5 to calculate p-values. Extreme outliers that were larger than 75% quartile + 3 × (75% quartile − 25% quartile) were removed before plotting. | true | true | true |
|
PMC9603433 | Gladys Y.-P. Ko,Fei Yu,Kayla J. Bayless,Michael L. Ko | MicroRNA-150 (miR-150) and Diabetic Retinopathy: Is miR-150 Only a Biomarker or Does It Contribute to Disease Progression? | 11-10-2022 | microRNA,diabetes,oxidative stress,inflammation,apoptosis,pathological angiogenesis,retinopathy | Diabetic retinopathy (DR) is a chronic disease associated with diabetes mellitus and is a leading cause of visual impairment among the working population in the US. Clinically, DR has been diagnosed and treated as a vascular complication, but it adversely impacts both neural retina and retinal vasculature. Degeneration of retinal neurons and microvasculature manifests in the diabetic retina and early stages of DR. Retinal photoreceptors undergo apoptosis shortly after the onset of diabetes, which contributes to the retinal dysfunction and microvascular complications leading to vision impairment. Chronic inflammation is a hallmark of diabetes and a contributor to cell apoptosis, and retinal photoreceptors are a major source of intraocular inflammation that contributes to vascular abnormalities in diabetes. As the levels of microRNAs (miRs) are changed in the plasma and vitreous of diabetic patients, miRs have been suggested as biomarkers to determine the progression of diabetic ocular diseases, including DR. However, few miRs have been thoroughly investigated as contributors to the pathogenesis of DR. Among these miRs, miR-150 is downregulated in diabetic patients and is an endogenous suppressor of inflammation, apoptosis, and pathological angiogenesis. In this review, how miR-150 and its downstream targets contribute to diabetes-associated retinal degeneration and pathological angiogenesis in DR are discussed. Currently, there is no effective treatment to stop or reverse diabetes-caused neural and vascular degeneration in the retina. Understanding the molecular mechanism of the pathogenesis of DR may shed light for the future development of more effective treatments for DR and other diabetes-associated ocular diseases. | MicroRNA-150 (miR-150) and Diabetic Retinopathy: Is miR-150 Only a Biomarker or Does It Contribute to Disease Progression?
Diabetic retinopathy (DR) is a chronic disease associated with diabetes mellitus and is a leading cause of visual impairment among the working population in the US. Clinically, DR has been diagnosed and treated as a vascular complication, but it adversely impacts both neural retina and retinal vasculature. Degeneration of retinal neurons and microvasculature manifests in the diabetic retina and early stages of DR. Retinal photoreceptors undergo apoptosis shortly after the onset of diabetes, which contributes to the retinal dysfunction and microvascular complications leading to vision impairment. Chronic inflammation is a hallmark of diabetes and a contributor to cell apoptosis, and retinal photoreceptors are a major source of intraocular inflammation that contributes to vascular abnormalities in diabetes. As the levels of microRNAs (miRs) are changed in the plasma and vitreous of diabetic patients, miRs have been suggested as biomarkers to determine the progression of diabetic ocular diseases, including DR. However, few miRs have been thoroughly investigated as contributors to the pathogenesis of DR. Among these miRs, miR-150 is downregulated in diabetic patients and is an endogenous suppressor of inflammation, apoptosis, and pathological angiogenesis. In this review, how miR-150 and its downstream targets contribute to diabetes-associated retinal degeneration and pathological angiogenesis in DR are discussed. Currently, there is no effective treatment to stop or reverse diabetes-caused neural and vascular degeneration in the retina. Understanding the molecular mechanism of the pathogenesis of DR may shed light for the future development of more effective treatments for DR and other diabetes-associated ocular diseases.
Diabetes is a disease characterized by hyperglycemia associated with either insulin deficiency or resistance, and the incidence of diabetes is projected to increase to 33% of the US population by 2050 owing to the obesity epidemic [1], of which 90–95% of diabetic patients will have type 2 diabetes (T2D) [2]. Diabetic retinopathy (DR) is a chronic complication associated with both T1D and T2D. It impacts 4.2 million people in the US and 93 million worldwide [3] and is a leading cause of blindness among the working population in the US [4]. Overall, DR is diagnosed in 30% of diabetic patients: approximately 90% of T1D and 60% of T2D patients develop DR [5]. The risk factors for developing DR include the duration of diabetes (≥20 years), poor control over blood glucose levels, hypertension, and obesity [6]. Clinically, DR has been diagnosed and treated as a vascular disease, but it also affects the neural retina [7]. Diabetic insults impair the integrity of retinal microvasculature and induces pathological angiogenesis [8]. Depending on the severity of the vascular pathologies, DR is divided into non-proliferative and proliferative phases clinically. Non-proliferative DR (NPDR) manifests mild-to-moderate vascular abnormalities including microaneurysms, intraretinal hemorrhages, and venous beading, while proliferative DR (PDR) displays neovascularization and pre-retinal hemorrhages [5]. In addition, the decreased retinal light responses recorded by electroretinogram (ERG) correlate with more severe vascular pathologies in NPDR patients [9]. As chronic diabetic conditions adversely impact the neural retina and retinal vasculature, the interaction between retinal neurons and vascular cells could further contribute to the pathogenesis of DR. Laser photocoagulation is a commonly used therapy for PDR but is invasive and often induces blind spots in the retina [5]. The most used therapy for DR is the intraocular injections of anti-vascular endothelial growth factor (VEGF) agents [10]. However, nearly 30% of patients do not respond well to anti-VEGFs [11,12], and less than 50% of patients have improved vision after 1–2 years of anti-VEGF therapies [13]. As repeated anti-VEGF treatments are needed to conquer the recurrent neovascularization, they often cause unwanted side effects, including retinal detachment [13]. In addition, current therapies for DR mainly target neovascularization at the later stages of DR and rarely restore normal visual function [5]. Therefore, it is critical to understand the mechanisms underlying the pathogenesis of DR and develop therapeutic strategies to either target the early stage of DR or to prevent its development.
Neural dysfunction and degeneration occur early in the diabetic retina. Distorted color vision in patients with early diabetes was previously reported [14,15,16], and dysfunction of the neural retina can be detected in diabetic patients by ERG before any vascular pathologies are detected [17,18,19]. Diabetic patients without DR vascular complications usually have lower ERG amplitudes and longer implicit times than healthy subjects [20]. Apoptotic non-vascular cells can be found in the retina of diabetic patients before the onset of DR [21]. In T1D patients without retinopathy, the thickness of the retinal nerve fiber layer (NFL) decreases compared with healthy subjects, suggesting the loss of axons from retinal ganglion cells (RGCs) [22], and the dampened vision correlates with thinner neural layers [23]. While most apoptotic neurons are found in the retinal ganglion cell layer (GCL), the outer nuclear layer (ONL; photoreceptors) also displays apoptosis [24]. In patients with six years of diabetes duration, the apoptotic cells increase in the retina from the ONL to GCL [25]. The apoptotic markers are detected in the retina of diabetic patients, including caspase-3, a protease for apoptosis, in the GCL and Fas ligand (FasL) in major retinal layers (from NFL to ONL) [26]. Adolescents with T2D for an average of two years have reduced retinal thickness measured by optical coherence tomography (OCT) and dampened light responses measured by ERG [27]. In T2D patients without retinopathy, the thickness of retinal layers from GCL to the outer plexiform layer (OPL) decrease after one year of follow-up [28]. In T2D patients with mild NPDR (microaneurysms), the thicknesses of retinal NFL, GCL, and inner plexiform layer (IPL) decrease, indicating neurodegeneration in the initial stage of DR [29,30]. The decreased thickness of NFL correlates with the severity of DR in T2D patients, suggesting that neurodegeneration in the diabetic retina might exacerbate the development of DR [31]. In diabetic animal models, decreased thickness of the retinal inner nuclear layer (INL) and IPL occur in the retina of T1D mice (Ins2Atita) with increased expression of caspase-3 [32]. The number of cells decreases while the expression of caspase-3 increases in the GCL after ten weeks of diabetes in streptozotocin (STZ)-induced T1D mice [33]. In a mouse model of T2D (KKAY), the terminal UTP nick-end label (TUNEL) staining shows an increased number of apoptotic neurons in the GCL [34], and neuronal apoptosis in the retina occurs in T2D (db/db) mice starting from twenty weeks old [35]. These findings reveal the degeneration of the inner retina in diabetes, which may explain the dampened light response reflected by the decreased amplitude and increased implicit time of ERG b-waves [36,37]. Among retinal neurons, photoreceptors undergo apoptosis shortly after the onset of diabetes [38,39]. The ERG a-waves from diabetic patients [40] and animals [41] display decreased amplitudes and increased implicit times compared with healthy counterparts, which indicate a diabetes-induced impairment to the photoreceptors. In T2D patients, decreased thicknesses of the ONL and the inner and outer segments of photoreceptors are associated with the development of retinopathy [28]. In patients with metabolic syndromes, the thickness of the photoreceptor layer decreases compared with healthy subjects [42]. Apoptotic photoreceptors can be detected in STZ rats 4 weeks after the onset of diabetes [38], while the thickness of the ONL decreases in STZ mice after 10 weeks of diabetes [33]. Electron microscopy shows disorganization and degeneration of the outer segments of photoreceptors in STZ rats [43]. In addition, 28-week-old T2D mice (db/db) have decreased thickness of the ONL accompanied by dampened light responses on ERG [35]. Interestingly, long-term diabetic patients with retinitis pigmentosa (RP), a genetic disease with loss of photoreceptors, rarely develop DR even though these patients develop other diabetes-related vascular diseases [44,45]. In a mouse model of RP, during the period when photoreceptors are undergoing apoptosis, the retinal vasculature is also degenerating. Once the photoreceptors are completely lost, the vascular degeneration stops [46]. Hence, photoreceptor apoptosis not only contributes to the neural dysfunction under diabetes but may also adversely impact diabetic microvascular complications [46]. However, how diabetic insults cause photoreceptor apoptosis remains unclear.
The TUNEL labelling of microvascular networks in trypsin-digested retinas from diabetic patients shows significantly increased apoptotic endothelial cells and pericytes compared with healthy subjects [47,48]. The loss of pericytes and the formation of acellular capillaries are the major signs of microvascular degeneration that occurs at an earlier stage of DR [7]. Acellular capillaries contain only the basement membrane and remnants of endothelial cells without nuclei, which are typical pathological changes found in trypsin-digested diabetic retinas [49]. Another sign of microvascular degeneration is the loss of pericytes or the existence of “ghost” pericytes that appear as light-stained pockets around the basement membrane [50]. Degenerated vessels are detected in the eyes of patients with mild NPDR and may contribute to the formation of microaneurysms [51]. In alloxan-induced T1D rats, the number of apoptotic vascular cells and acellular capillaries are increased in trypsin-digested retinas [52]. Apoptotic endothelial cells are also increased in the retina of STZ-T1D rats, but inhibition of FasL dampens apoptosis [53]. Apoptotic pericytes expressing the pro-apoptotic BCL2 associated X (BAX) protein, an apoptosis regulator, can be detected in the retina of diabetic patients [54]. The TUNEL-positive pericytes are detected in the retina of STZ-T1D rats [55], while a loss of pericytes is found in db/db-T2D mice [56]. The number of acellular capillaries and ghost pericytes are increased in the retina of STZ-T1D and Zucker-T2D rats along with elevated activities of caspase-3, and inhibition of tumor necrosis factor (TNF-α), a pro-inflammatory cytokine that can trigger necrosis or apoptosis, alleviates those pathological changes [57]. Microvascular degeneration may contribute to the breakdown of the blood–retina barrier (BRB) [58] and exacerbate the decrease of blood flow and local hypoxia in the diabetic retina [59]. Retinal hypoxia stimulates the secretion of angiogenic factors such as VEGF from various cell types in the retina, including astrocytes and Müller glia [60]. The upregulated VEGF eventually leads to pathological angiogenesis and neovascularization in DR.
Oxidative stress is due to overproduction or decreased removal of reactive oxygen species (ROS) in cells. Mitochondria are the major organelle that produce OS during oxidative phosphorylation [61]. Normally, the electrons donated by reduced nicotinamide adenine dinucleotide (NADH) and flavin adenine dinucleotide (FADH2) are transferred from complex I to complex IV in the inner mitochondrial membrane by coenzyme Q and cytochrome C. Meanwhile, protons are transferred to the intermembrane space to generate a gradient of protons between the mitochondrial matrix and the intermembrane space. The influx of protons drives the synthesis of ATP while the electrons are consumed to produce H2O [62]. Diabetes-associated hyperglycemic conditions promote the generation of electron donors NADH and FADH2, which pushes the proton gradient to the threshold and ultimately hinders the transfer of electrons. Alternatively, electrons are provided to O2 by coenzyme Q to generate superoxide and ROS [63]. Nicotinamide adenine dinucleotide phosphate (NADPH) also donates electrons to O2 through the membrane proteins NADPH oxidases (Nox) [64]. Under diabetic conditions, protein kinase C (PKC) is activated that further increases the activity of Nox [65]. The Nox proteins are highly expressed in the vasculature, so ROS generated from Nox contributes to impaired vascular function [66]. The increased ROS in mitochondria promotes the opening of mitochondrial permeability transition pores (mPTP) and increases the release of cytochrome C [67], which eventually induces apoptosis. Eight-hydroxydeoxyguanosine (8-OHdG) is a biomarker for oxidative stress-induced DNA damage which is increased in the vitreous of T2D patients indicating upregulated oxidative stress in the retina [68]. Antioxidant reagents have been used to alleviate the diabetes-caused apoptosis of retinal neurons and endothelial cells [61]. Glutathione (GSH) is a major endogenous antioxidant that removes ROS and suppresses oxidative stress. In STZ-induced diabetic mice, GSH is decreased in the mitochondria of the retina, which correlates to the increase of degenerated (acellular) retinal capillaries [69]. Antioxidant α-lipoic acid treatment decreases the apoptotic microvascular cells after 11 months of STZ-induced diabetes, and it decreases 8-OHdG and increases GSH in the retina [70]. Treatment with the antioxidant lutein reduces the activity of caspase-3, an enzyme leading to apoptosis, and alleviates STZ-induced neural dysfunction [71].
Inflammation is a hallmark of diabetes that manifests in the diabetic retina [72]. A functional blood–retina barrier (BRB) in the diabetic retina protects the neural retina from the invasion of immune cells in the circulation [73]. Before the breakdown of the BRB in the advanced stages of DR, the retinal glial cells, neurons, and endothelial cells are the major sources of diabetes-induced inflammation [74,75,76]. The innate immune system in the retina responds to diabetic insults by activating microglia and secreting pro-inflammatory molecules. Retinal microglia are the resident immune cells derived from monocytes. In addition, the circulating macrophages can differentiate into microglia upon stimulation by low expression of CD45. Resting retinal microglia reside in the inner and outer plexiform layers with ramified shapes, while the activated microglia change to an amoeboid shape and migrate to various retinal layers [73]. In mouse retinas under oxidative stress, the apoptosis of photoreceptors increases concurrently with the activation of microglia in the photoreceptor layers. Activated microglia secrete pro-inflammatory factors, including TNF-α and interleukins (ILs), which exacerbate inflammatory reactions and promote apoptosis in vascular cells and neurons [77]. In cultured human retinal endothelial cells (HRECs), treatments of IL-1β and TNF-α increase the activities of caspase-3 and caspase-8 [78], activate the pro-inflammatory nuclear factor kappa B (NFĸB), and upregulate the expressions of intracellular adhesion molecule (ICAM)-1 and vascular cell adhesion molecule (VCAM)-1 [79]. Increased activities of caspase-3 and -8 indicate an increase of cell apoptosis, and upregulations of ICAM-1 and VCAM-1 induce leukostasis (adherence of leukocytes) in the retinal vessels and promote further inflammatory reactions by mediating the migration of leukocytes [80]. Knocking out the type 1 interleukin-1 receptor (IL1R1) in STZ-diabetic mice largely decreases the activities of the caspases and the number of acellular capillaries [81], and inhibition of TNF-α decreases the activities of caspase-3 and -8 and the apoptosis of RECs in STZ-induced diabetic rats. Furthermore, diabetic mice with deficient TNF-α receptors (TNFR1 and TNFR2) have decreased acellular capillaries and increased pericytes compared with wild-type diabetic mice [82]. These data clearly demonstrate that inhibition of pro-inflammatory factors or their receptors will decrease apoptosis of endothelial cells and preserve the retinal microvasculature. Increased IL-1β and TNF-α also induce apoptosis in retinal neurons. Under hypoxia, the expressions of IL-1β and TNF-α in retinal microglia and the corresponding receptors IL-1R1 and TNFR1 in retinal ganglion cells (RGCs) are increased, which leads to apoptosis of RGCs. Neutralizing IL-1β and TNF-α with antibodies suppresses the apoptosis of RGCs [83]. In the retinal degeneration 1 (rd1) mouse model, blocking the downstream signaling of interleukin-1 receptors (IL-1R) alleviates the degeneration of photoreceptors and improves the light responses of the retina [84]. Thus, diabetes-elicited secretion of pro-inflammatory molecules from microglia triggers apoptosis in the neural retina. The astrocytes and Müller glia cells also contribute to the inflammation in the diabetic retina. Under hyperglycemia, astrocytes have increased activation of NFĸB and production of ROS as well as elevated levels of pro-inflammatory factors including IL-1β, TNF-α, and monocyte chemoattractant protein-1 (MCP-1) [85]. The MCP-1 may further recruit and activate microglia, which accelerate the local inflammatory response [86]. The Müller glial cells in the retina of STZ-induced diabetic rats have increased expression of pro-inflammatory factors, such as ICAM-1 [87]. Increased IL-1β in the diabetic retina can stimulate the expression of IL-6 and activate NFĸB in Müller cells, suggesting that the Müller cells mediate the exacerbation of inflammation in the diabetic retina [88]. In addition, the production and secretion of anti-inflammatory pigment epithelium-derived factor (PEDF) [89] are decreased in Müller cells under hyperglycemic [90] or hypoxic [91] conditions, but the expression of pro-inflammatory VEGF in Müller cells is increased under these conditions [92,93]. Taken together, microglia, astrocytes, and Müller cells are all involved in diabetes-elicited inflammation and contribute to apoptosis in the neural retina.
One mechanism of neuronal apoptosis in the diabetic retina is glutamate excitotoxicity. Glutamate, an excitatory neurotransmitter, mediates the synaptic transmission in the retina [94]. Uptake of glutamate in the synaptic cleft by neurons and glial cells is necessary to maintain the concentration of extracellular glutamate and cease the activation of the postsynaptic receptors [95]. In the diabetic retina, glutamate [96] and its receptors are upregulated [97]. In addition, the activity of glutamate transporters is reduced in Müller glia cells under diabetes [98,99]. Increased glutamate release and reduced glutamate transporters induce extended activation of the glutamate receptors allowing excessive influx of calcium into neurons [100]. The elevated intracellular calcium is transported into the mitochondrial matrix and activates the PTP, which facilitates the release of cytochrome C and production of ROS and leads to the apoptosis of neurons [101]. Hence, diabetes-elicited changes in glutamate, its receptors, and glutamate transporters in the retina cause glutamate toxicity that also contributes to neuronal apoptosis.
MicroRNAs (miRs) are short, non-coding, single-stranded RNAs approximately 23 nucleotides in size, and they target one or more downstream messenger RNAs (mRNAs) causing post-transcriptional degradation or translational repression [102,103,104]. The mature miR is derived from a precursor sequence transcribed from the genome by either RNA polymerase II or III, and miR expression shows tissue- and developmental-stage-specific patterns [102,103]. MicroRNAs inhibit the translation of target mRNAs by preventing the initiation of translation [105,106,107,108], inhibiting the elongation of translation [109,110,111], and inducing the degradation of target mRNAs [102,103,112]. MicroRNAs represent a set of modulators that regulate metabolism, inflammation, and angiogenesis [113], and they have also been linked to DR [114,115]. Changes of miR levels in various organs or blood have been reported in diabetic patients and animals, and in particular, changes in circulating or retinal miRs correlate to some disease progressions and have been suggested as biomarkers for chronic diseases associated with diabetes including DR [114,116]. In STZ-induced diabetic rats, at least 86 miRs are altered in the retina [115,117,118]. Patients with T1D have circulating miR-29a, miR-148a, miR-181a, and miR-200a upregulated, while miR-21a, miR-93, miR-126, and miR-146a are downregulated [119]. The level of miR-126 negatively correlates with the risk of developing PDR [120]. The decreased miR-150 and increased miR-30b detected from the plasma of T1D patients are associated with the development of DR [121]. In the plasma of T2D patients and obese-hyperglycemic mice (ob/ob), the levels of miR-15a, miR-20b, miR-21, miR-24, miR-126, miR-191, miR-197, miR-320, miR-486, and miR-150 are decreased [122]. The downregulation of miR-20b in the serum of T2D patients correlates with the development of DR and may be used to predict the severity of DR [123]. While diabetes-associated changes of miRs clearly demonstrate a correlation between miRs and DR progression, few miRs have been shown to directly contribute to the pathogenesis of DR.
As described in previous sections, inflammation, oxidative stress, and cell apoptosis are part of the pathogenesis of DR, and changed miRs in the blood or retina could be involved in these processes [124]. In cultured retinal ganglion cells (RGCs) treated with a high concentration of glucose (HG), the expression of miR-495 is increased. Overexpression of miR-495 further exacerbates the HG-induced RGC apoptosis, while its inhibition protects RGCs against cell death [125]. In the retina of high-fat-diet-induced T2D rats, retinal miR-93-5p is decreased. Overexpression of miR-93-5p in the diabetic retina alleviates the microvascular degeneration, downregulates pro-inflammatory factors (IL-1β, IL-6, and TNF-α), and elevates antioxidant levels, including GSH and superoxide dismutase (SOD) [126]. MiR-21 is decreased in the retina of T2D mice (db/db) and in RECs treated with palmitic acid, a condition mimicking an extracellular high-fat environment. Knocking out miR-21 in T2D mice alleviates the degeneration and leukostasis of the retinal microvasculature, decreases the levels of pro-inflammatory factors (TNF-α and VCAM-1), and upregulates the antioxidant PPARα in the retina [127]. Overexpression of miR-145 in cultured RECs alleviates the HG-induced apoptosis in part because the activation of toll-like receptor 4 (TLR4) mediates inflammatory responses and promotes the activation of NFĸB [128], and TLR4 is a downstream target of miR-145. In HG-treated RECs, overexpression of miR-145 suppresses the expression of TLR4 and inhibits the activation of NFĸB and production of other pro-inflammatory factors (IL-1β and TNF-α) [129]. In STZ-diabetic rats, miR-195 is increased in the retinal GCL, INL, ONL, and RECs compared with non-diabetic rats. The upregulation of miR-195 also occurs in cultured RECs treated with HG. The expression of manganese superoxide dismutase (MnSOD), an endogenous antioxidant, is decreased in the STZ-diabetic retina and cultured RECs treated with HG, which correlates with increased oxidative stress and apoptosis [130,131]. Inhibition of miR-195 mitigates the STZ-diabetes- and HG-induced suppression of MnSOD, thus alleviates diabetes and HG-elicited apoptosis [132]. MiR-146a is downregulated in the circulation of T2D patients [133] and also decreased in cultured RECs treated with HG [134]. Decreased miR-146a correlates with escalated inflammation [135]. In STZ-diabetic rats, intraocular injection of miR-146a suppresses the diabetes-induced increase of the pro-inflammatory intercellular adhesion molecule 1 (ICAM1) and mitigates the damage to retinal light response and microvascular integrity [118]. Overexpression of miR-146a inhibits the inflammatory response in HG-treated RECs by suppressing the expressions of TLR4, phosphorylated NFĸB, and TNF-α and blocking the downstream signaling of TLR4 [134]. MiR-15a is decreased in the RECs of diabetic patients compared with non-diabetic subjects. Overexpression of miR-15a in the mouse retina inhibits the expression of pro-inflammatory factors, including IL-1β, IL-6, and TNF-α [136]. MiR-20b is decreased in the serum of T2D patients compared with healthy subjects, and T2DR patients have further decreases of miR-20b compared with T2D patients without retinopathy [123]. Overexpression of miR-20b-3p in the eyes of STZ-diabetic rats alleviates the visual dysfunction as well as neural and vascular degeneration in the retina by reducing BAX but increasing BCL-2, thus reducing apoptosis in the retina of STZ-diabetic rats. In addition, the expressions of pro-inflammatory factors (IL-1β and TNF-α) are downregulated by overexpressing miR-20b-3p in the diabetic retina [137]. These examples further demonstrate the crucial roles of miRs in the development of DR. (Table 1).
MicroRNA-150 is downregulated in patients with obesity [138], T1D [139,140], and T2D [121]. In high-fat-diet (HFD)-induced T2D mice, miR-150 is decreased in the plasma and retina [141,142]. Downregulation of miR-150 is also observed in the heart of STZ-diabetic rats [143] and in the ischemic retina of mice [144]. Inhibition of miR-150 promotes apoptosis [145], while its overexpression alleviates the apoptosis of cells under hypoxia [146], in which local hypoxia occurs in the diabetic retina [147,148]. Overexpression of miR-150 also protects the retinal vasculature from degeneration induced by oxygen-induced retinopathy, a model for hypoxia-induced angiogenesis [149]. Moreover, miR-150 is an intrinsic suppressor of inflammation [113]. Overexpression of miR-150 downregulates TNF-α and NFĸB induced by lipopolysaccharide (LPS) in endothelial cells [150]. Deletion of miR-150 (miR-150−/−) exacerbates the increase of IL-1β, IL-6, and TNF-α in mice with HFD-induced T2D [113]. We observed that plasma and retinal miR-150 is decreased in mice fed with an HFD even before diabetes develops [142]. The miR-150 knockout (miR-150−/−) mice with HFD-induced T2D display more severe retinal neural dysfunction and vascular pathologies compared with wild-type (WT) mice with HFD-T2D (Figure 1A) [141,142]. Therefore, decreased miR-150 correlates with the development of diabetes and may facilitate the development of DR by promoting apoptosis and inflammation in the neural and vascular retina.
MicroRNAs often have many targets, and a single mRNA can also be targeted by multiple miRs [102,103,151], and the biological processes mediated by miRs and their targets are often tissue- and cell-type-specific [103,152]. Decreased miR-150 may promote apoptosis and inflammation in the diabetic retina through upregulating its downstream target genes. There are confirmed target genes of miR-150 that can regulate inflammation. In HFD-induced T2D mice, decreased miR-150 upregulates its target genes MYB proto-oncogene (Myb), ETS-domain transcription factor 1 (Elk1), and eukaryotic translation termination factor 1 (Etf1). Knocking down Myb, Elk1, or Etf1 suppresses the inflammatory response by inhibiting the activation of B cells [113]. Early growth response 1 (Egr1) is another target gene of miR-150 [153], and knockdown of this target alleviates the diabetes-induced inflammation in mouse mesangial cells by downregulating pro-inflammatory factors (IL-1β, IL-6, and TNF-α) [154,155]. Moreover, these target genes of miR-150 (Myb, Elk1, Etf1, and Egr1) are also involved in the regulation of apoptosis. Knocking out Myb upregulates the apoptosis of mouse colorectal carcinoma cells [156], and overexpressing Myb decreases the production of ROS and alleviates the apoptosis in cardiomyocytes after hypoxia/reoxygenation injury [157]. Overexpression of ELK1 protein has been found to induce apoptosis in neurons by interacting with the mitochondrial permeability transition pore complex (PTP) [158]. Transfection of Elk1 in the dendrites of primary neurons induces apoptosis [159], while inhibition of Elk1 alleviates the apoptosis of neurons under oxygen–glucose deprivation [160]. Upregulated Etf1 is associated with decreased apoptosis in mouse pre-osteoblast cells [161]. Increased expression of Egr1 is associated with the apoptosis of squamous cell carcinoma cells and breast cancer cells, while knocking down Egr1 mitigates apoptosis [162,163]. In addition to Myb and Egr2 (direct targets of miR-150) that are known to promote angiogenesis by increasing the population of hemogenic endothelial cells [164] or upregulating vascular endothelial growth factor (VEGF) and its receptor 2 (VEGFR2) expressions [165,166], respectively, both VEGF and VEGFR2 are downstream of miR-150 [141,149]. VEGF and its principal receptor for angiogenesis VEGFR2 are both upregulated in diabetic eyes [167,168,169], and anti-VEGF therapies have been used to treat neovascularization in DR [169,170,171]. As miR-150 is downregulated in diabetes and DR, its downstream targets are upregulated and correlate with diabetes-associated inflammation, oxidative stress, apoptosis, and pathological angiogenesis. Thus, miR-150 is not only a biomarker for diabetes and DR, but it could also be a potential therapeutic target for treating diabetes-associated chronic diseases and DR.
As mentioned previously, inflammation is a major contributor to DR [72,172]. Chronic meta-inflammation is a hallmark of obesity and obesity-associated type 2 diabetes (T2D) [173,174], but numerous studies have indicated that intraocular rather than systemic inflammation is more closely associated with the vascular complications in DR [46,74,175,176,177,178,179,180,181,182,183,184]. Interestingly, diabetic patients who also have retinitis pigmentosa (RP), a congenital blindness with initial degeneration of rod photoreceptors, rarely develop DR [185,186,187], and there is a clear inverse correlation between RP and DR [185,186]. The RP patients who had been diabetic for nearly 40 years developed other non-retinal vascular complications, but none had retinal microaneurysms, exudates, or any clinical DR [185,186,187]. In mice, genetic deletion of rod photoreceptors or pharmacological inhibition of photoreceptors reduces retinal inflammation and alleviates progression of DR [188,189]. Therefore, retinal photoreceptors are a major source of intraocular inflammation and directly contribute to vascular abnormalities in diabetes [6,22,23,24,33,34,36]. MiR-150 is an intrinsic suppressor of inflammation [114] since it suppresses the expression of pro-inflammatory molecules and cytokines, including nuclear factor kappa B (NF-ĸB), TNFα, IL1β, and IL6 [113,190,191,192,193,194]. Overexpression of miR-150 downregulates lipopolysaccharide (LPS)-induced expression of TNF-α and NF-ĸB in endothelial cells [150], while deletion of miR-150 in mice (miR-150−/−) augments LPS-stimulated inflammatory responses [113]. MiR-150−/− mice with HFD-induced T2D have further elevated serum pro-inflammatory cytokines (TNFα, IL1β, IL6, CCL2) and lower anti-inflammatory cytokine (IL10) [113]. Compared with wild-type (WT) mice with HFD-induced T2D, these miR-150−/−-T2D mice display more severe T2D with increased glucose intolerance and insulin resistance [113] and significantly reduced retinal light responses [141,142]. Thus, miR-150 exhibits anti-inflammatory [113] properties, and diabetes-associated decrease of miR-150 may contribute to ocular inflammation and further exacerbate the development of DR. We previously showed that miR-150−/−-T2D mice have more severe inflammation in photoreceptors and exacerbated vascular degeneration compared with the WT-HFD mice [142]. Since the biological processes mediated by microRNAs and their targets are often tissue- and cell-type-dependent as stated earlier [152,195], after screening the top 30 predicted target genes of miR-150 [113,193] and identifying new bona fide targets that are pro-inflammatory [113], multiple transcription factors including the eukaryotic translation termination factor 1 (Etf1), early growth response 1 (Egr1), MYB proto-oncogene (Myb), and ETS-domain transcription factor 1 (Elk1) were found to be expressed in retinal photoreceptors and endothelial cells [196]. Downregulation of miR-150 correlates with an upregulation of Etf1, Egr1, Myb, and Elk1 and pro-inflammatory cytokines, while overexpression of miR-150 or knocking down any of these transfection factors decrease the expression of pro-inflammatory cytokines in cultured adipose B lymphocytes [113]. In the diabetic retina, photoreceptors are one of the major sources of retinal inflammation [175,189]. Using cultured murine photoreceptors treated with palmitic acid (PA) to mimic obesity-associated T2D, we found that PA elicited an increase of phosphorylated NF-ĸB (pP65), an inflammation marker, which persisted for 24 h and correlated with a persisting decrease of miR-150 and increase of Elk1. However, PA elicited only temporary increases of Etf1, Egr1, or Myb, although these three transcription factors are direct downstream targets of miR-150 and expressed in the neural retina [196]. Overexpression of miR-150 or knocking down Elk1 not only decreased the expression of ELK1 (protein) but also relieved PA-induced increase of inflammation. Phosphorylated ELK1 at S383 (pELK1S383) translocates from the cytoplasm to the nucleus at which time it then activates its downstream genes to promote inflammation [197,198]. The level of pELK1S383 was increased in the retinal outer nuclear layer of obesity-associated T2D mice and the nuclei of palmitic acid-treated murine photoreceptors in cultures, and the increased nuclear pELK1S383 correlated with the upregulated pP65. Deletion of miR-150 not only upregulated ELK1 but also cytoplasmic pELK1S383 in photoreceptors. The miR-150−/− mice with obesity-associated T2DR had further exacerbated retinal photoreceptor inflammation compared with the WT-T2DR mice, and the photoreceptor inflammation correlated with an increase of pELK1S383 in the retinal outer nuclear layer [196]. The nuclear/cytoplasmic (N/C) ratio represents the cytoplasm-to-nucleus translocation of pELK1S383, and PA treatments increased the N/C ratio of pELK1S383 in cultured photoreceptors, and knocking down Elk1 decreased nuclear pELK1S383 and the N/C ratio of pELK1S383 in PA-treated photoreceptors, which also correlated with the downregulation of pP65. Hence, T2D-associated inflammation in photoreceptors was in part mediated by a decrease of miR-150 that caused an increase of nuclear pELK1S383 and led to photoreceptor inflammation. Therefore, overexpression of miR-150 or knocking down of Elk1 may restrain the development of DR by mitigating the inflammation in the neural retina, especially in photoreceptors [196].
The neural retina has the highest oxygen consumption rate among all tissues, including the brain [199], thus making it (especially the photoreceptors) prone to hypoxia-induced apoptosis. There is local hypoxia in the diabetic retina, and among retinal neurons, photoreceptors undergo apoptosis shortly after the onset of diabetes [38,39]. Patients with T1D have thinner neural layers in the retina and visual dysfunction before the diagnostics of DR [23]. Adolescents with T2D for an average of two years have reduced retinal thickness and dampened light responses [27]. The loss of retinal neurons starts 10 weeks after STZ-induced T1D in mice [33], and neuronal apoptosis in the retina occurs in T2D (db/db) mice from 20 weeks of age [35]. Apoptotic photoreceptors can be detected in STZ-diabetic rats 4 weeks after the onset of diabetes [38]. In addition, the dysfunction of photoreceptors in STZ-diabetic mice is associated with the reduced thickness of the outer nuclear layer [200]. Furthermore, diabetic patients with retinitis pigmentosa (RP), a genetic disease with loss of photoreceptors, rarely develop DR, even though these patients develop other diabetes-related vascular diseases [44,45]. In a mouse model of RP, during the period when photoreceptors are undergoing apoptosis, the retinal vasculature is also degenerating. Once all photoreceptors have died, the vascular degeneration stops [46]. Hence, photoreceptor apoptosis in diabetes not only contributes to the neural dysfunction but may also adversely impact diabetic microvascular complications [46,200] and lead to DR. Subsets of miRs are known to regulate cell proliferation and apoptosis [139]. Among them, inhibition of miR-150 promotes apoptosis of cells under hypoxia [145], and local hypoxia occurs in the early diabetic retina [147,148]. Overexpression of miR-150 alleviates the apoptosis of hypoxic cells [146]. In our HFD/obesity-associated T2D mouse model, not only did HFD-T2D mice have more severe retinal neural dysfunction and apoptotic photoreceptors versus mice fed with a normal diet [141,142], the miR-150−/− mice with HFD-T2D had even more neural dysfunction and the highest numbers of apoptotic photoreceptors compared with the WT mice with HFD-T2D [196]. To further verify the relationship between miR-150 and photoreceptor apoptosis, we treated cultured photoreceptors with PA to elicit cell apoptosis. Interestingly, knocking down miR-150 in photoreceptors caused a significant increase in apoptosis regardless of PA treatments. However, transfections with miR-150 mimics did attenuate the PA-induced apoptosis. Thus, overexpression of miR-150 only in photoreceptors alone might not be enough to overturn PA-induced apoptosis, but an adequate level of miR-150 is necessary for the survival of photoreceptors [196]. Among the major targets of miR-150 expressed in photoreceptors, Elk1 is known to promote apoptosis in neurons [159,201], as overexpression of ELK1 protein has been found to promote apoptosis in neurons. Transfection of Elk1 in the dendrites of primary neurons induces apoptosis [159], while inhibition of Elk1 alleviates the apoptosis of neurons under oxygen–glucose deprivation [160]. As mentioned previously, the activation of ELK1 requires its phosphorylation, and phosphorylation of ELK1 at threonine 417 (pELK1T417) specifically is essential for ELK1-mediated neuronal apoptosis [158]. Thus, we analyzed the levels of ELK1 and pELK1T417 in the inner and outer segments of photoreceptors (IS + OS) as well as in the outer nuclear layer (ONL) from the retinas of HFD-T2D mice [196]. We found that the levels of ELK1 in the cytoplasm (IS + OS) and nuclei (ONL) of photoreceptors were significantly increased in both WT and miR-150−/− mice with HFD-T2D. Knockout of miR-150 (miR-150−/−) upregulated pELK1T417 in the IS + OS of photoreceptors, while increased pELK1T417 was observed in the ONL of all HFD-T2D retinas. It is possible that HFD-induced apoptosis in photoreceptors was mediated by an increase in nuclear pELK1T417, and the upregulation of cytoplasmic pELK1T417 caused by miR-150 knockout exacerbated the HFD-induced apoptosis. To further determine the relationship of miR-150, ELK1/pELK1T417 and photoreceptor apoptosis, we employed PA-induced apoptosis in cultured photoreceptors [196]. After photoreceptors were treated with PA, the levels of ELK1 significantly increased in a time-dependent manner, which correlated with increased apoptosis. As treatments with PA (24 h) significantly increased ELK1 in all cells, knocking down miR-150 further elevated the PA-elicited increase in ELK1. However, overexpression of miR-150 did not attenuate the PA-induced increase of ELK1 suggesting that overexpression of miR-150 is not sufficient to downregulate PA-stimulated ELK1, which echoes that overexpression of miR-150 in photoreceptors alone might not be enough to overturn PA-induced apoptosis. In order to verify the functions of ELK1 and pELK1T417 in regulating apoptosis in photoreceptors, we knocked down Elk1 with siRNA (siElk1) in cultured photoreceptors and found that PA-elicited increase of ELK1 was blocked by siElk1, and knocking down Elk1 decreased cytoplasmic pELK1T417 and also arrested the PA-induced increase in nuclear pELK1T417. Thus, knocking down Elk1 effectively inhibited the PA-elicited increase in ELK1 and pELK1T417. Unfortunately, PA-induced apoptosis was not dampened by siElk1, so knockdown of Elk1 alone cannot attenuate PA-induced apoptosis, which is consistent with our data where upregulation of miR-150 in photoreceptors alone was not enough to conquer PA-induced apoptosis [196]. Cell apoptosis can be mediated by the mitochondrial permeability transition pore complex (PTP) that initiates mitochondrial swelling and membrane potential depolarization that leads to cell death [202]. There is a protein–protein interaction between ELK1 and PTP in the brain, and cytoplasmic ELK1 can be isolated from purified mitochondrial fractions. Furthermore, cell apoptosis induced by Elk1 overexpression can be blocked by a PTP inhibitor in cultured primary neurons [201]. Thus, under T2D conditions, upregulated cytoplasmic pELK1T417 would have increased interactions with mitochondrial PTP, which might further accelerate photoreceptor apoptosis. However, while overexpression of miR-150 or downregulation of Elk1 decreases cytoplasmic pELK1T417, it does not reduce nuclear pELK1T417 or overcome PA-elicited apoptosis. In PA-treated photoreceptors, the nuclear/cytoplasmic (N/C) ratio of pELK1T417 remains comparable with cells with/without overexpression of miR-150 and cells with/without knockdown of Elk1. The N/C ratio represents the cytoplasm-to-nucleus translocation of pELK1T417, which is important for trans-activating the downstream targets of Elk1 and regulating apoptosis [203,204]. The translocation of pELK1T417 to the cell nucleus correlates with increased apoptosis in neurons [205]. Therefore, in addition to dampening the expression of ELK1, blocking the translocation of pELK1T417 to the nucleus may be more critical to mitigate diabetes-associated apoptosis in photoreceptors [196].
As neural miR-150 is important for protecting the retina under diabetic insults (described above), decreased miR-150 in vascular endothelial cells may also contribute to ocular angiogenesis. Overexpression of miR-150 in mouse eyes protects the retinal microvasculature from degeneration induced by oxygen-induced retinopathy, a model for pathological angiogenesis in retinopathy of prematurity [205,206], and deletion of miR-150 exacerbates T2D-associated microvascular leakage and degeneration [141,142]. Since well-known diabetic mouse models do not have pathological neovascularization like in PDR patients, we used a three-dimensional collagen matrix culture system of endothelial cells to evaluate the effect of endothelial miR-150 in vascular sprouting, an indicator of neovascularization (Figure 1). We found that overexpression of miR-150 dampened endothelial sprouting and invasion, but inhibition of miR-150 did not affect normal endothelial sprouting (Figure 1B). Furthermore, overexpression of miR-150 decreased the expression of VEGFR2 in cultured endothelial cells [141]. Interestingly, VEGFR2 is not a direct downstream target of miR-150, since the VEGFR2 gene lacks compatible paired sequences. Conversely, the miRNAs predicted to target the 3′-UTR (1479 bp) of the mouse VEGFR2 gene do not include miR-150 [141]. One direct target of miR-150 that may regulate the expression of VEGFR2 is the transcription factor Myb, which binds to a 5′-YAACKG-3′ sequence in the promoter region and regulates the expression of a group of genes involved in cell lineage- and fate-determination in the immune system. The gene encoding VEGFR2 (Vegfr2) has four Myb binding sites in its promoter region, so Vegfr2 can be turned on by Myb [67]. Overexpression of Myb increases the population of hemogenic endothelial cells during embryonic development [68]. We found that overexpression of miR-150 also decreased the expression of MYB in cultured endothelial cells [141], making Myb the most likely downstream target of miR-150 that regulates VEGFR2 expression in vascular endothelial cells. Table 2 is a list of the direct targets of miR-150 discussed above.
As decreased miR-150 in the diabetic retina correlates with the development of DR, the action and downstream targets of miR-150 in neural versus vascular retina are different, but all contribute to the pathogenesis of DR (Figure 2). In the neural retina, diabetes-associated decrease of miR-150 promotes inflammation and apoptosis of photoreceptors via Elk1, which contribute to the microvascular degeneration in DR. In the retina vasculature, diabetes-associated decrease of miR-150 promotes endothelial cell sprouting via Myb, which contributes to neovascularization in DR (Figure 2). MiR-150 is expressed in the neural and vascular retina and is also abundant in the circulation, so diabetes-elicited changes in circulating miR-150 are reported in both T1D and T2D patients. Hence, miR-150 is not only a biomarker for DR, it is indeed involved in the pathogenesis and the disease progression of DR. With extensive investigation of miR-150 as an example, diabetes-associated changes in other miRs not only serve as biomarkers to indicate the pathological progression of DR, but they might also actively contribute to the pathogenesis of DR. | true | true | true |
|
PMC9603556 | Xin Li,Hao Zhang,Yong Wang,Yanyan Li,Youli Wang,Jiangjiang Zhu,Yaqiu Lin | Chi-Circ_0006511 Positively Regulates the Differentiation of Goat Intramuscular Adipocytes via Novel-miR-87/CD36 Axis | 14-10-2022 | goat,circRNA,adipocyte differentiation,novel-miR-87,CD36 | Goats are an important livestock and goat meat is essential to local people. The intramuscular fat (IMF) content has a great influence on the quality of goat meat. The intramuscular preadipocytes differentiation is closely related to the IMF deposition; however, its potential regulatory mechanisms remain unclear. CircRNAs were revealed to be involved in multiple biological progressions. In this study, we took primary goat intramuscular preadipocyte (GIMPA) as the study model to verify the function and mechanism of chi-circ_0006511, which was abundant and up-regulated in mature adipocytes (GIMA). The results showed that the expression level of chi-circ_0006511 gradually increased in the early stage of GIMPA differentiation, and chi-circ_0006511 was confirmed to promote GIMPA lipid droplets aggregation and up-regulate the adipogenic differentiation determinants, further promoting GIMPA differentiation. Mechanistically, chi-circ_0006511 exerts its function by sponging novel-miR-87, thereby regulating the expression of CD36. The results from this study provided novel significant information to better understand the molecular regulatory mechanism of intramuscular preadipocytes differentiation, thereby providing a new reference for the intramuscular fat adipogenesis in goats. | Chi-Circ_0006511 Positively Regulates the Differentiation of Goat Intramuscular Adipocytes via Novel-miR-87/CD36 Axis
Goats are an important livestock and goat meat is essential to local people. The intramuscular fat (IMF) content has a great influence on the quality of goat meat. The intramuscular preadipocytes differentiation is closely related to the IMF deposition; however, its potential regulatory mechanisms remain unclear. CircRNAs were revealed to be involved in multiple biological progressions. In this study, we took primary goat intramuscular preadipocyte (GIMPA) as the study model to verify the function and mechanism of chi-circ_0006511, which was abundant and up-regulated in mature adipocytes (GIMA). The results showed that the expression level of chi-circ_0006511 gradually increased in the early stage of GIMPA differentiation, and chi-circ_0006511 was confirmed to promote GIMPA lipid droplets aggregation and up-regulate the adipogenic differentiation determinants, further promoting GIMPA differentiation. Mechanistically, chi-circ_0006511 exerts its function by sponging novel-miR-87, thereby regulating the expression of CD36. The results from this study provided novel significant information to better understand the molecular regulatory mechanism of intramuscular preadipocytes differentiation, thereby providing a new reference for the intramuscular fat adipogenesis in goats.
Goats are the most widespread of all domesticated ruminants due to their extreme adaptability [1]. Consumers are increasingly interested in goat meat as these animals have a small amount of subcutaneous and intramuscular fat [2]. The potential of goats to produce high-quality meat is mainly reflected in their healthy fats, low-calorie intramuscular fats, saturated fats, and especially their high proportions of unsaturated and saturated fatty acids [3].In addition, appearance, tenderness, taste and juiciness are important categories that influence consumer acceptance of goat meat [4,5].The content of intramuscular fat (IMF) is one of the key factors in determining the tenderness and juiciness of meat, and also greatly affects the flavor [6,7]. The study of goat IMF deposition is of great significance for improving meat quality and breeding. While IMF deposition was mainly determined by the hyperplasia and hypertrophy of intramuscular adipocytes (IMA) [8], adipocyte differentiation is an important part of IMF deposition, which is an extremely complex physiological process that is tightly regulated by multiple transcription factors and noncoding RNAs(ncRNA) [9,10,11,12]. PPARγ and C/EBPα play key roles in adipogenic differentiation as classic adipogenic differentiation marker genes [13,14,15]. C/EBPβ can activate C/EBPα and PPARγ and produce a cascade reaction that rapidly activates the expression of adipogenesis-related genes [16,17,18,19]; C/EBPα is translated in the early stages of adipocyte differentiation and cooperates with PPARγ to promote adipocyte differentiation [20]. Furthermore, LPL, SREBP1, RXRα, and some members of the KLF family are also involved in the regulation of adipocyte differentiation [10]. ncRNAs, including microRNAs (miRNAs), long non-coding RNAs (lncRNAs) and circular RNAs (circRNAs), have also been shown to be important regulators of adipocyte differentiation in recent years [21,22,23]. For instance, miR-340-5p inhibits ovine adipocyte differentiation by targeting ATF7 [24] and miR-25-3p regulates the differentiation of intramuscular preadipocytes in goats via targeting KLF4 [25]. He et al. (2022) previously constructed an lncRNA-miRNA-mRNA regulatory network during goat intramuscular and subcutaneous adipocyte differentiation by RNA-seq, and identified several lncRNAs that may regulate adipocyte differentiation [26]. Though circRNAs have been reported to affect adipocyte differentiation in cattle [27], pigs [28], and ducks [29], there have been no reports on circRNAs regulating adipocyte differentiation in goats. Therefore, exploring new endogenous regulatory factors is of substantial significance for elucidating the differentiation in goat intramuscular adipocytes. As a kind of endogenous ncRNA, circRNA widely exists in various cells of various organisms, is abundantly expressed, and has high cell specificity, tissue specificity, and developmental stage specificity [30,31]. CircRNA is produced by the covalent attachment of spliceosome-mediated mRNA 3’ splice sites to 5’ splice sites [32,33]. CircRNA was first recorded in plant viroids in 1976 [34]. According to the origin and generation pattern of genome, circRNA can be divided into three types: intronic circRNAs [35], exonic circRNAs [36], and exon–intron circRNAs [37].CircRNAs have a variety of biological functions, including as a competing endogenous RNA (ceRNA) to sponge miRNA [38], regulate the transcription of its host genes [39], and circRNA also has the potential to encode proteins [40,41,42]. Among them, circRNAs are the most widely studied as ceRNA mechanisms. Based on the multiple biological functions of circRNAs, circRNAs are extensively involved in biological processes such as the occurrence and development of cancer, osteogenic/myogenic/adipogenic differentiation, lipid metabolism, and browning of white adipose tissue [43,44,45,46,47,48]. A large number of studies have shown that circRNAs function as ceRNA to regulate adipocyte differentiation. For instance, in cattle, it has been reported that circFUT10 inhibits adipocyte differentiation via sponging let-7 [45], and circPPARγ facilitates adipocyte differentiation by binding miR-92a-3p and YinYang 1 [49]. However, the effect of circRNAs on goat adipocyte differentiation is still unknown. In order to identify the circRNAs function in goat adipocyte, the differentially expressed chi_circ_0006511 screened by whole transcriptome sequencing (RNA-seq) of goat intramuscular adipocytes during differentiation was selected as the research object. In this study, we used fluorescence in situ hybridization (FISH), nucleocytoplasmic separation, RNA pull down, overexpression, knockdown, Bodipy and Oil red O staining, dual luciferase report assay and qPCR to explore the effect of chi-circ_0006511 on the differentiation of goat intramuscular adipocytes, and confirmed that the chi-circ_0006511/novel-miR-87/CD36 axis positively regulates GIMPA differentiation, which not only enriched the regulatory network of goat intramuscular adipocyte differentiation but also provides new ideas for further study of the function of fat deposition circRNA. These results provide a basis for exploring the molecular mechanism of intramuscular adipocyte differentiation and serve as a reference for molecular breeding in goats.
We obtained 16 differentially expressed circRNAs (data in submission) through the RNA-seq of GIMPA and GIMA (induced for differentiation for 3 days) (Figure 1A). In this study, one of the up-regulated circRNAs, chi_circ_0006511, was selected as the target. chi_circ_0006511 was formed by the circularization of exon 9 and exon 11 of the LMO7 gene on goat chromosome 22, which was a typical exonic circRNA (Figure 1B). Divergent primers containing splicing sites were designed according to the sequence and Convergent primers were designed according to the linear transcript of the chi_circ_0006511 host gene. Using cDNA reverse transcribed from total RNA and linear RNA digested with RNase R as templates, PCR was performed with PrimeSTAR® Max DNA Polymerase, bands were verified by electrophoresis (Figure 1C), and the presence of splice sites was verified by Sanger sequencing (Figure 1B).
We detected the expression of chi-circ_0006511 at different stages of GIMPA differentiation, and the results showed that chi-circ_0006511 was up-regulated at the early stage of differentiation, and then decreased (Figure 2A), suggesting that it has a promoting effect to GIMPA differentiation. Therefore, the constructed pcDNA3.1-circ_0006511-EF1-ZsGreen was transfected as OE group, pcDNA3.1-null-EF1-ZsGreen was transfected as control group (OE-NC), and the transfection efficiency of OE and OE-NC was observed by green fluorescence (Figure 2B). The expression of chi-circ_0006511 increased about 40-fold (Figure 2C), and more lipid droplets accumulation in the OE group was observed by Bodipy and Oil red O staining (Figure 2D), and differentiation marker genes, PPARγ, C/EBPα, C/EBPβ, LPL, SREBP1 were extremely significantly up-regulated (p < 0.01) (Figure 2E), indicating that overexpression of chi-circ_0006511 promotes the differentiation of goat intramuscular adipocytes. To further verify the effect of chi-circ_0006511 on GIMPA differentiation, we used chemically synthesized siRNA to knockdown the expression of chi-circ_0006511 and found that the efficiency of the first siRNA(si1) was about 70% (Figure 3A); therefore, si1 was used for subsequent experiments. We observed that compared with the siNC group, the number of lipid droplets in the si1 group was reduced (Figure 3B) and the differentiation marker genes were extremely significantly down-regulated (p < 0.01) (Figure 3C), which was contrary to the result of overexpressing chi-circ_0006511, indicating that chi-circ_0006511 was a positive regulator of GIMPA differentiation.
The intracellular distribution of circRNA is closely related to its function. In order to explore how chi-circ_0006511 regulates GIMPA differentiation, we first identified the subcellular localization of chi-circ_0006511 by nucleocytoplasmic separation and FISH. The results showed that chi-circ_0006511 was distributed in both the nucleus and cytoplasm of goat intramuscular adipocytes (Figure 4A,B), suggesting that it may function as ceRNA. Therefore, we used the circMir, a MiRanda and RNAHybrid-based program to predict its possible sponged miRNAs, obtaining 10 common miRNAs. Theoretically, these miRNAs have opposite expression trends with circRNAs; since chi-circ_0006511 was an up-regulated circRNA in GIMA, we selected 141 down-regulated miRNAs (DEmir-down) in GIMA from our previous RNA-seq data. Combined with circMir results, five common miRNAs were obtained (Figure 4C,D). Then, according to the minimum free energy (mfe) of miRNAs binding with circRNA predicted by RNAHybrid, novel-miR-87 was selected for further study (Figure 4D). Subsequently, the expression of novel-miR-87 after overexpression and knockdown of chi-circ_0006511 were detected, and it was found that chi-circ_0006511 negatively regulates the expression of novel-miR-87(Figure 4E,F), suggesting that it may be a potential target miRNA of chi-circ_0006511. To further verify this speculation, we then designed a biotin-labeled specific probe for chi-circ_0006511 and confirmed that novel-miR-87 was pulled down together with chi-circ_0006511 by RNA pull down (Figure 4G). Finally, the dual-luciferase reporter assay showed that novel-miR-87 mimics inhibited the dual-luciferase activity of psiCHECK2-circ_0006511 WT. In summary, these results indicate that chi-circ_0006511 acts as a ceRNA sponge novel-miR-87.
What role does novel-miR-87 play in GIMPA differentiation? We synthesized the mimics/inhibitor based on the sequence of novel-miR-87 to mimic/inhibit its expression. The results showed that novel-miR-87 mimics were up-regulated by about 12,000-fold (Figure 5A). At the same time, it was found that lipid droplet accumulation was reduced compared with the mim-NC group (Figure 5B), and adipocyte differentiation marker genes were significantly down-regulated (p < 0.01, Figure 5C). As expected, novel-miR-87 inhibitor treatment promoted GIMPA differentiation (Figure 5D–F), indicating that novel-miR-87 was a negative regulator of GIMPA differentiation, which was opposite to the effect of chi-circ_0006511.
miRNAs mainly regulate the expression of target genes by binding to the 3’UTR region of target genes. Therefore, we used RNAHybrid to predict the target genes of novel-miR-87 and compared them with the target genes of novel-miR-87 predicted by previous transcriptome sequencing. By intersection, 1209 shared genes were obtained (Figure 6A). Then we performed KEGG analysis on 1209 genes and displayed the pathways related to adipocyte differentiation. CD36, which was simultaneously enriched in fat digestion and absorption, the PPARg signaling pathway, adipocytokine signaling pathway and the AMPK signaling pathway, was selected as a potential target gene of novel-miR-87 (Figure 6B). The expression of CD36 was opposite to that of novel-miR-87, novel-miR-87 mimics extremely significantly inhibited CD36 expression (p < 0.01, Figure 6C), while novel-miR-87 inhibitor extremely significantly up-regulated CD36 expression (p < 0.01, Figure 6D). On the other hand, CD36 was consistent with the expression changes in chi-circ_0006511, and was significantly increased/decreased with overexpression/knockdown of chi-circ_0006511 (p < 0.01) (Figure 6E,F). Additionally, novel-miR-87 mimics significantly inhibited the dual-luciferase activity of psiCHECK2-CD36 3’WT (p < 0.01, Figure 6G).
In order to reveal the effect of CD36 on GIMPA differentiation, we constructed the px459-CD36 vector based on CRISPR/cas9 to knockdown CD36. Firstly, two sgRNAs were designed by CRISPOR (tefor.net) based on the CD36 sequence. The constructed vector was verified by monoclonal PCR and sanger sequencing to confirm that the sgRNA was successfully inserted into the vector (Figure 7A). The successfully constructed vector was then transferred into GIMPA, and the control group (named CD36-KDNC) was transfected with px459 plasmid, CD36-KD1 containing sg1, and CD36-KD2 containing sg2. The expression of CD36 was detected by qPCR and WB (Figure 7B), the results showed that CD36-KD1 had a high knockout efficiency. Compared with CD36-KDNC, less lipid droplets and down-regulated differentiation marker genes was observed in CD36-KD1 (both p < 0.01, Figure 7C,D). These results indicated that suppressed CD36 inhibited the differentiation of goat intramuscular adipocytes.
circRNAs had a different splicing feature and special loop structures; the identification of back splicing sites was the basis of circRNAs function verification. For this reason, before elucidating the role of chi_circ_0006511 in regulating GIMPA differentiation, its loop formation needs to be identified. The chi_circ_0006511 back-splice site identified by RNase R digestion, PCR and sanger sequencing was consistent with the sequence by RNA-seq, which proved that the chi_circ_0006511 actually circulated in GIMA. Based on this, we found that chi_circ_0006511 showed an upward trend from 0 to 84 h of GIMPA differentiation, and a down-regulated trend from 84 to 120 h. It was speculated that chi_circ_0006511 might play a positive role in the early differentiation of GIMPA. Functionally, the overexpression of chi_circ_0006511 promoted lipid accumulation in GIMA, and the mRNA expression levels of adipogenic genes PPARγ, C/EBPα, C/EBPβ, SREBP1, LPL were significantly up-regulated. The upregulation of the expression of these genes suggested that the gain-of-function of chi-circ_0006511 promotes the differentiation of GIMPA. In contrast, knockdown of chi_circ_0006511 by siRNA inhibited lipid accumulation in GIMA, and the mRNA expression levels of adipogenic genes were significantly down-regulated, which was opposite to the trend of overexpression. It could be concluded that chi_circ_0006511 promoted GIMPA differentiation. Numerous studies have reported that circRNAs act as miRNA sponges to regulate fat deposition in animals. For instance, Jiang et al. (2020) revealed that circFUT10 was abundantly expressed in Qinchuan bovine subadipocytes and circFUT10 combined with let-7c promotes cell proliferation and inhibits cell differentiation by targeting PPARGC1B in bovine adipocytes. Li et al. (2022) reported that circPPARA affects porcine IMF content via adsorbed miR-429 and miR-200b [50]. chi_circ_0006511 belong to exonic circRNA. The composition of circRNAs was closely related to the subcellular localization of circRNAs, which in turn was closely related to the biological functions of circRNAs. Among the many mechanisms of circRNA, the most widely studied and most concerned is the ceRNA mechanism. Exonic RNAs are mostly distributed in the cytoplasm and usually have one or more miRNA binding sites, which can absorb miRNAs by sponge, thereby releasing the inhibitory effect of miRNAs on target genes [51]. Therefore, we first examined the subcellular localization of chi_circ_0006511 and found that chi_circ_0006511 was expressed in both the nucleus and cytoplasm of GIMA, suggesting that it may function as a molecular sponge for miRNAs. By predicting the miRNA bound by chi_circ_0006511 and selecting novel-miR-87 for verification, it was found that chi_circ_0006511 negatively regulates the expression of novel-miR-87, and further confirmed by RNA pull down and dual luciferase reporter assay that chi_circ_0006511 sponged novel-miR-87. Additionally, the inhibitory effect of novel-miR-87 on GIMPA differentiation was verified by gain-of-function and loss-of-function experiments, which was opposite to that of chi_circ_0006511. miRNAs bind to the 3’ UTR region of target genes, inhibit the expression of target genes, and play an important regulatory role in animal growth and development, cell proliferation, cell differentiation and apoptosis [52,53]. Then, the target genes of novel-miR-87 were predicted by RNAHybrid and RNA-seq, and 1209 common genes were obtained. KEGG analysis of these common genes showed that the PI3K/Akt signaling pathway, adipocytokine signaling pathway, AMPK signaling pathway and other pathways were significantly enriched. Among these, the PI3K/Akt signaling pathway was known to play a central role in cell physiology by mediating growth factor signaling, glucose homeostasis, lipid metabolism, and cell proliferation [54]. In adipocytes, the PI3K/Akt signaling pathway promotes lipid biosynthesis and inhibits lipid hydrolysis and is a positive regulator of adipocyte differentiation [55]. The adipocytokine signaling pathway is the sum of all proteins and factors responsible for regulating adipocytokine, mainly including APN, leptin, IL-6 and TNF-α and so on. APN inhibits adipogenesis through the AMPK pathway, which regulates lipid, cholesterol, and glucose metabolism in specialized metabolic tissues such as liver, muscle, and adipose tissue [56]. CD36 was commonly enriched in the PPARg signaling pathway, adipocytokine signaling pathway and AMPK signaling pathway, and studies have reported that CircScd1 inhibits the formation of lipid droplets in AML-12 hepatocytes through the JAK2/STAT5 pathway, which is achieved by inhibiting the expression of CD36 to block its mediated lipid uptake [57,58,59]. Therefore, we selected CD36 as a potential target gene of novel-miR-87 for further study and verified that novel-miR-87 targeting CD36 3’UTR negatively regulates the expression of CD36. It was also found that CD36 was up-regulated with the overexpression of chi-circ_0006511, and down-regulated with knockdown of chi-circ_0006511. In order to elucidate the regulatory effect of CD36 on the differentiation of goat intramuscular adipocytes, the goat px459-CD36 vector for knockdown CD36 was constructed, and vectors were transferred to GIMPAs. Oil red O and Bodipy staining showed that knockdown of CD36 inhibited the accumulation of lipid droplets, and the expression levels of adipogenic differentiation marker genes PPARγ, C/EBPα, C/EBPβ and LPL were significantly down-regulated. LPL is mainly expressed in tissues with large amounts of oxidized or stored fatty acids, such as heart, skeletal muscle, brown adipose tissue and white adipose tissue [60]. LPL acts as a scavenging factor lipase [61], can hydrolyze the triglyceride-rich lipoprotein VLDL and triglycerides in chylomicrons. Loss-function of LPL mice were born with significantly elevated blood TG levels and die from inability to absorb butterfat [62], suggesting that LPL is particularly important for fatty acid uptake [63]. Long-chain fatty acids (LCFAs) provide energy for cells, and are also an important component of cell membranes and intracellular lipid storage materials [64]. CD36, a type of membrane protein, is a member of the class B scavenger receptor family and is also a fatty acid transmembrane transporter. CD36 uptakes LCFAs in tissues and promotes lipid accumulation and dyslipidemia [65]. Adipose tissue-specific KO-CD36 mice has reduced cold tolerance because CD36 knockout inhibits lipid accumulation in brown adipose tissue, hinders the uptake of LCFAs in adipose tissue, and increases the content of free LCFAs in blood [66]. As a sensor of LCFAs, CD36 can also activate the PPAR signaling pathway [67] and AMPK signaling pathway [68], while PPARγ can induce CD36 uptake of lipids [69]. The differentiation of GIMA was inhibited with CD36 knockdown, which may be due to the inhibition of the uptake and transport capacity of LCFAs by CD36 and LPL down-regulated, thereby reducing the energy intake during the differentiation of GIMA. At the same time, due to the knockdown of CD36, the activation of the PPARγ and AMPK signaling pathways was inhibited. Our previous research found that CD36 was widely expressed in goat tissues, with the highest expression in adipose tissue, and CD36 expression was positively correlated with goat IMF content, and this study showed that knockdown of CD36 inhibited the differentiation of goat intramuscular adipocytes, indicating that CD36 had a positive regulation on goat IMF deposition. However, although the regulatory role of CD36 in the differentiation of goat intramuscular adipocytes was not yet clear, the specific regulatory mechanism still needs to be further studied.
Goat intramuscular preadipocytes (GIMPA) were isolated from the longissimus dorsi muscle (Sichuan jianyang dageda animal husbandry CO., LTD, Sichuan, China) of 7-day-old Jianzhou Da-er goats. The isolation and culture of cells was consistent with the previous study [25]. DMEM-F12 (Gibco, Carlsbad, CA, USA) containing 10% FBS (fetal bovine serum, Gibco, USA), 1‰ Penicillin-Streptomycin (Gibco, Carlsbad, CA, USA) and 50 μmol·L−1 oleic acid (Sigma, St Louis, MO, USA) was used to induce adipogenic differentiation of GIMPA to goat intramuscular adipocytes (GIMA).
Total RNA was extracted by Trizol reagent (TaKaRa, Tokyo, Japan) according to the instructions. Total RNA (2 μg) was incubated with 3 U/μg of RNase R (Geneseed, Guangzhou, China) for 15 min at 37 °C. After treatment with RNase R, the RNA was equally divided into two parts, one for electrophoresis verification and one for cDNA synthesis by the Revert Aid First Stand cDNA Synthesis Kit (Thermo FisherScientific, Waltham, MA, USA) according to the instructions; the expression levels of chi_circ_0006511 and mRNAs were analyzed by qRT-PCR. Then, 1 μg of untreated RNA was taken as a template directly for reverse transcription; RNA and cDNA were stored at −80 °C and −20 °C, respectively.
cDNA was used as the template and reaction system: 10 μL TB GreenTMPremix Ex TaqTMII, 1 μL Foward and reverse primers, 1 μL cDNA, 7 μL ddH2O, PCR program: 95 °C for 3 min; 95 °C for 30 s, annealing for 15 s, extension at 72 °C for 15 s, 40 cycles. The primer information is shown in Table 1. qPCR results were processed by the 2−ΔΔCt method, with UXT as the internal reference gene, and the data were expressed as “Mean ± SD”. One-way ANOVA in SPSS 24.0 was used for significance analysis, and Graphpad prism 9.0 was used for drawing.
The specific probe of chi_circ_0006511 conjugated with Cy3 was designed and synthesized by RiboBio CO., LTD (Guangzhou, China). The cellular localization of chi_circ_0006511 was detected by FISH kit (RiboBio, Guangzhou, China). Briefly, the GIMAs cultured with slides in 24 wells were fixed with 4% paraformaldehyde, permeabilized with pre-cooled PBS containing 0.5% Triton X-100 for 10 min, added with prehybridization solution and incubated at 37 °C for 30 min, and then 50 μmol of probe was added to the pre-warmed hybridization solution and incubated with cells at 37 °C overnight. In dark conditions, the slides were washed with 4× (0.1% Tween-20), 2× and 1 × SSC solutions at 42 °C, respectively, and then the slides were fixed on the slides with Mounting Medium with DAPI (abcam, Cambridge, UK). Pictures were taken with a confocal microscope (Zeiss, Oberkochen, Germany).
Cytoplasmic and nuclear fractions of 1 × 107 GIMAs were isolated by the PARIS™ Kit (AM1556, Thermo Fisher Scientific, Bothell, WA, USA) according to the introduction. Briefly, GIMAs was lysed in Cell Fraction Buffer on ice for 10 min. After centrifugation at 500× g for 3 min at 4 °C, the supernatant was collected as a cytoplasmic fraction, followed by washing the pellet with Cell Fraction Buffer. Finally, the nuclei were collected.
chi_circ_0006511 overexpression vector (pcDNA3.1-circ_0006511-EF1-ZsGreen) and control vector (pcDNA3.1-null-EF1-ZsGreen) were obtained from HanBio (Shanghai, China). The Crispr/cas9 CD36 vector based on the Px459 for knocking down the expression level of CD36 was constructed following the protocol of Zhang Lab. The sgRNA of CD36 was designed using CRISPOR software, two pairs of sgRNAs with higher scores were annealed and phosphorylated, then the Quick Ligase (NEB, Beijing, China) was used to connect it with the px459 plasmid purified by Bbs I digestion (Thermo Fisher Scientific, WA, USA); after picking a single clone, it was verified by sequencing that the inserted sgRNA was correct. Goat chi_circ_0006511 wild-type (psiCHECK2- chi_circ_0006511 WT) and mutant (psiCHECK2- chi_circ_0006511 MT) vectors for reporting dual-luciferase activity were constructed by Tsingke (Chengdu, China). CD36 3′UTR was cloned by 3′RACE kit (Takara, Tokyo, Japan), and the CD36 3′ wild-type (psiCHECK2-CD36 3′ WT) and mutant (psiCHECK2-CD36 3′ MT) were constructed by Tsingke (Chengdu, China).
The interfering RNA of Chi_circ_0006511 was designed and synthesized by HanBio (Shanghai, China) (Table 2). According to the sequence of novel-miR-87, novel-miR-87 mimics and novel-miR-87 inhibitor were designed and synthesized by GemePharma (Shanghai, China) (Table 3). TurboFect™ Transfection Reagent (Invitrogen, Carlsbad, CA, USA) was used to transfect vector, siRNA, miRNA mimics and miRNA inhibitor into 80% confluent GIMPAs, according to the manufacturer’s instructions.
The luciferase activity was detected after co-transfection of psiCHECK2-chi_circ_0006511 and psiCHECK2-CD36 3’ with novel-miR-87 mimics, respectively. The dual luciferase reporter system (Vazyme, Nanjing, China) was used to detect the binding relation between chi_circ_0006511 and chi_novel-miR-87, chi_novel-miR-87 and CD36.
The cells were washed three times by PBS. After discarding the PBS, the cells were fixed with 4% paraformaldehyde for 30 min, and the formaldehyde was discarded and then washed twice with PBS. The Bodipy stock solution (Thermo fisher Scientific, Waltham, MA, USA) was diluted 1:1000 with PBS under dark conditions to obtain the working solution, and 200 μL of the working solution was added to each well for 30 min at room temperature, and the lipids in the goat intramuscular adipocytes were observed by fluorescence microscope. The Oil red O stock solution (Solarbio, Beijing, China) and ddH2O were mixed at a ratio of 3:2 and filtered twice to obtain the Oil red O working solution. Then, 200 μL Oil Red O working solution was added to each well for 30 min at room temperature in the dark for staining, the Oil red O dye was discarded, washed three times with PBS and lipid droplet accumulation was observed and photographed in GIMAs.
The equal amount of protein lysis was separated by 10% SDS-PAGE, and then the protein was transferred to PVDF membrane (Millipore, Billerica, MA, USA), blocked with 5% skim milk at room temperature for 2 h, washed with TBST, incubated with primary antibody overnight at 4°C, and then incubated with secondary antibody. After one hour of incubation at room temperature, the immunoreactive signals were visualized by an ECL kit (BioRad, Hercules, CA, USA). Anti-CD36 for western blot (1:500 dilution, Wanleibio, Shenyang, China), Goat Anti-Rabbit IgG (H + L) HRP for western blot (1:5000 dilution, Abways, Shanghai, China), β-Actin Mouse Monoclonal Antibody for western blot (1:1000 dilution, Abways, Shanghai, China).
RNA pull down was performed to demonstrate the binding between circRNA and miRNA by PureBindingTM RNA-Protein pull-down Kit (Geneseed, Guangzhou, China). Briefly, chi-circ_0006511 specific probes conjugated with biotin (RiboBio, Guangzhou, China) bind to magnetic beads carrying streptavidin, incubated with cell lysate. Then, the magnetic beads were collected using a magnetic frame, and chi-circ_0006511 was pulled down. The obtained total RNA was reverse transcribed using the Mir-X miRNA First-Strand Synthesis Kit (Takara, Tokyo, Japan) and miRNA was detected by qPCR.
In conclusion, this study clarified the positive regulatory effect of chi-circ_0006511 on GIMPA differentiation and identified mechanistically that chi-circ_0006511 acts through the novel-miR-87/CD36 axis (Figure 8). | true | true | true |
|
PMC9603599 | Chuantian Xie,Zhaojun Ding | NAC1 Maintains Root Meristem Activity by Repressing the Transcription of E2Fa in Arabidopsis | 14-10-2022 | root,meristem,NAC1,E2Fa,endoreduplication | Root meristem is a reserve of undifferentiated cells which guide root development. To maintain root meristem identity and therefore continuous root growth, the rate of cell differentiation must coordinate with the rate of generation of new cells. The E2 promoter-binding factor a (E2Fa) has been shown to regulate root growth through controlling G1/S cell cycle transitions in Arabidopsis thaliana. Here, we found that NAC1, a member of the NAM/ATAF/CUC family of transcription factors, regulated root growth by directly repressing the transcription of E2Fa. Loss of NAC1 triggers an up-regulation of the E2Fa expression and causes a reduced meristem size and short-root phenotype, which are largely rescued by mutation of E2Fa. Further analysis showed that NAC1 was shown to regulate root meristem by controlling endopolyploidy levels in an E2Fa-dependent manner. This study provides evidence to show that NAC1 maintains root meristem size and root growth by directly repressing the transcription of E2Fa in Arabidopsis. | NAC1 Maintains Root Meristem Activity by Repressing the Transcription of E2Fa in Arabidopsis
Root meristem is a reserve of undifferentiated cells which guide root development. To maintain root meristem identity and therefore continuous root growth, the rate of cell differentiation must coordinate with the rate of generation of new cells. The E2 promoter-binding factor a (E2Fa) has been shown to regulate root growth through controlling G1/S cell cycle transitions in Arabidopsis thaliana. Here, we found that NAC1, a member of the NAM/ATAF/CUC family of transcription factors, regulated root growth by directly repressing the transcription of E2Fa. Loss of NAC1 triggers an up-regulation of the E2Fa expression and causes a reduced meristem size and short-root phenotype, which are largely rescued by mutation of E2Fa. Further analysis showed that NAC1 was shown to regulate root meristem by controlling endopolyploidy levels in an E2Fa-dependent manner. This study provides evidence to show that NAC1 maintains root meristem size and root growth by directly repressing the transcription of E2Fa in Arabidopsis.
Plant meristem is a reserve of undifferentiated cells which guide postembryonic development [1,2,3,4]. In the root meristem of Arabidopsis, all root tissues originate from a stem cell niche which includes a small group of slowly dividing cells, quiescent center (QC), and the surrounding stem cells [5]. Stem cell daughters undergo a certain number of cell divisions until they reach the transition zone where divisions cease and cells start to differentiate [5,6]. To maintain root meristem and therefore continuous root growth, the rate of cell differentiation must coordinate with the rate of generation of new cells [1]. A delayed or accelerated transition to elongation causes either an increase or a decrease in root meristem, respectively [1,2]. The root meristem is located at the distal part of the growing root and continuously generates new cells [6]. Two sets of transcriptional routes have been shown to play a key role in the establishment and maintenance of root meristem activity through specifying the QC and stem cell identity [1]. One is the plant hormone auxin, which is transduced through AP2 transcription factors PLETHORAs (PLTs) to control root stem cell niche identity and root meristem activity [1,7]. The other one involves the GRAS family transcription factors SHORTROOT/SCARECROW (SHR/SCR) and their interacting proteins [8,9,10,11]. Mutating each of the key regulators causes the loss of the root QC and stem cell identity, eventually affecting root growth and development [1,12,13]. In addition, other plant hormones such as cytokinin (CK), gibberellin (GA), and brassinosteroids (BR) also play important roles in the maintenance of root meristem identity [14,15,16]. Eukaryotic E2 promoter-binding factors (E2Fs) are transcription factors that are major regulators of cell division, DNA repair, and cell differentiation [17,18]. Among the six E2F transcription factors (E2Fa, E2Fb, E2Fc, EDL1/E2Fe, EDL2/E2Fd, EDL3/E2Ff) in Arabidopsis [19], both E2Fa and E2Fb are transcriptional activators, while E2Fc acts as a repressor of the cell cycle [17,20,21]. The other three E2F proteins, repressing E2F-regulated reporter genes, are also functional repressors [18,22]. The E2Fa plays an important role in switching cell division to endoreduplication [23]. Overexpression of both E2Fa and its partner DIMERIZATION PARTNER (DPa) in Arabidopsis can stimulate cell proliferation and induce extra rounds of DNA replication, which result cell endoreduplication [17]. The E2Fa/DPa complex, controlling the expression of essential genes during the G1/S transition, can be dissociated by SUMOylation E3 METHYL METHANESULFONATE SENSITIVITY GENE21 (AtMMS21) [24]. In particular, the phenotype of defective root development in 35s::E2Fa-DPa transgenic seedlings is completely recovered when AtMMS21 is overexpressed [24]. Additionally, rapamycin (TOR) kinase is essential for the maintenance of mitotic activity by directly phosphorylating E2Fa and E2Fb in both the shoot and root apexes [25,26]. All these results suggest that E2Fa plays an important role in root growth and development. However, the underlying molecular mechanism regarding how E2Fa is finely regulated at the transcriptional level to ensure root meristem identity and root growth is not well understood. NAC (NAM, ATAF1/2, CUC) domain proteins are unique to plants and comprise large gene families [27]. There are around 105 NAC members which are involved in various aspects of plant development [27,28]. NAC transcription factors are characterized by a highly conserved NAM DNA-binding domain in the N-terminal region, accompanied by diverse C-terminal domains [29]. During lateral root initiation, NAC1 acts as a transcription activator to mediate auxin signaling, and microRNA (miRNA) 164 guides the cleavage of endogenous NAC1 mRNA in an auxin-dependent manner [30,31,32]. Furthermore, the NAC1 proteins are ubiquitinated and degraded by the 26S proteasome in a SINAT5 E3 controlled manner [31]. Novel functions are described in that NAC1 is induced in response to wounding and functions in promotion of root rip emergence [33]. However, the mechanistic studies of NAC1 on the functions of plant root growth are still unknown. In this study, we found that the NAC domain protein NAC1 plays an important role in root growth. NAC1 maintains root meristem size and therefore root growth through directly repressing the transcription of E2Fa. The stunted root growth and reduced root meristem size, which result from the decreased cell numbers in root meristem, were largely rescued by mutation of E2Fa in the nac1 mutant. This study not only expands our knowledge about the biological roles of NAC1, but it also improves our understanding of how the cell cycle regulator E2Fa is delicately regulated by NAC1 to maintain root meristem and root growth.
In contrast to the other members of the family whose expression is restricted to shoot meristem and flower (NAM, CUC2, NAP) or to vascular tissues (CmNACP), NAC1 is one NAC family member which is expressed in the root [32]. Although some expression can be detected in leaf primordia, wounding in the leaf explant, or in expanding cotyledons, the highest NAC1 expression is restricted to the lateral root initiation regions and root meristem [32,33]. The role of NAC1 in lateral root development has been highlighted by transducing auxin signals downstream of the F-box protein TIR1 [32], and whether NAC1 also plays a role in primary root growth was unknown. To test this hypothesis, two T-DNA insertion mutants, nac1–1 and nac1–2, were examined and showed defective root growth (Figure 1A). Considering that microRNAs (miRNAs) guised the cleavage of endogenous and transgenic NAC1 mRNA, we also generated overexpression transgenic plants expressing a cleavage-resistant form of NAC1 mRNA (mNAC1) according to the previous report [30]. Compared to wild-type plants, both nac1 mutants and nac1–1/nac1–2 F1 generation mutants had shorter roots, whereas the mNAC1 overexpression lines displayed longer roots (Figure 1A,B and Figure S1A,B). Similarly, the size of the root apical meristem and meristem cell number were reduced in the nac1, nac1–1/nac1–2 F1 generation mutants and increased in the mNAC1 overexpressing lines, whereas the cortical cell length in the root maturation zone was unaltered compared to the wild type (Figure 1C–F and Figure S1C–F). Likewise, root length, root apical meristem size, and meristem cell number were completely or largely restored in nac1–1/NAC1pro:NAC1 compared to nac1–1 (Figure 2A–F), confirming that these phenotypes were caused by a mutation in NAC1.
Considering the essential role of the E2Fa gene in controlling cell division and differentiation during the G1/S cell cycle [17,24,34], we first examined the expression of E2Fa by RT-qPCR since meristem cell numbers were decreased in the nac1 mutant (Figure 3A). We observed a higher transcript level of E2Fa in nac1–1, and a decrease in mNAC1ox-4 overexpression lines compared to wild type (Figure 3A). Consistently, the expression of E2Fapro:GUS, which is highly expressed in the meristem of wild-type roots, was dramatically accumulated in nac1–1 roots (Figure 3B,C). Furthermore, to examine whether the NAC1 affects the transcriptional activity of E2Fa in plants, we then used a transient transformation LUC reporter gene (E2Fapro:LUC) and performed transient expression assays in Arabidopsis leaf mesophyll protoplasts. The results showed that the LUC expression of the E2Fa promoter was specifically repressed by NAC1 compared with control (Figure 3D). Taken together, these results suggest that NAC1 inhibits the transcriptional activity of E2Fa.
To further investigate whether NAC1 directly binds to the E2Fa promoter in vivo, we performed a chromatin immunoprecipitation (ChIP) followed by a quantitative real-time PCR assay (RT-qPCR) using NAC1 overexpression seedlings (mNAC1ox-4). It was previously reported that the DNA-binding specificity of NAC proteins contain the core CGT[GA] [29]. The global genomic DNA-binding site analysis revealed that the E2Fa promoter contains the predicated transcriptional sites of NAC proteins which marked A–D and E negative controls (Figure 3E). Compared with the wild-type seedlings, the A–D but not E fragments of the E2Fa promoter were highly enriched in the mNAC1ox-4 seedlings, indicating that the NAC1 proteins were associated with the genomic regions of the E2Fa promoter in vivo (Figure 3E). Furthermore, the electrophoretic mobility shift assays (EMSA) showed that NAC1 bound to the E2Fa ChIP-positive fragment (P1 and P2) specifically in vitro, and this interaction could be abolished by adding specific competitor probes (Figure 3F). Consistently, using yeast one-hybrid assays, we observed the interaction between NAC1 and the E2Fa promoter in yeast cells, suggesting that NAC1 directly binds to the E2Fa promoter (Figure 3G). Considering that three Arabidopsis E2F proteins E2Fa, E2Fb, and E2Fc, but not other E2F factors, have been shown to interact with the retinoblastoma-related (RBR) proteins and play the same roles in G1/S transition [19,35,36], we also studied the relationship between NAC1 and E2Fb or E2Fc. Similarly, we performed transient expression assays in Arabidopsis leaf protoplasts and yeast one-hybrid assays to text whether NAC1 regulated E2Fb or E2Fc expression (Figure S2A–D). The results showed that NAC1 neither regulates the expression of E2Fb nor E2Fc, nor binds to their promoter directly in yeast one-hybrid assays, indicating NAC1 specifically inhibits E2Fa expression during root development.
To further investigate the role of NAC1-mediated inhibition of E2Fa in cell division and root growth, we crossed an e2fa mutant with the nac1–1 mutant. We found that several phenotypes of the nac1–1 mutant were largely recovered in the nac1–1/e2fa double mutant, including the shorter root, shorter root apical meristem, and reduced meristem cell number (Figure 4A–F). Significantly, co-overexpression E2Fa and DPa lines displayed a shorter root meristem and reduced root length compared with the WT [24] (Figure 5A–D). We also examined root meristem cell numbers in co-overexpression E2Fa and DPa lines. As shown in Figure 5A–D, root meristem cell numbers were reduced in the 35s::E2Fa-DPa seedlings but not in the E2Fa overexpressing lines, implying that the reduced root meristem size and root growth resulted from the decreased root meristem cell numbers in nac1–1. Taken together, these results indicated that E2Fa acts as a downstream signaling component of NAC1 to maintain root meristem size and root growth. The previous study showed that E2Fa-DPa was a key regulator of the endocycle [17]. To further study if the elevated E2Fa expression caused root growth arrest and might be a result of the regulation of endoreduplication in the nac1–1 mutant, we performed flow cytometric analysis in the WT, 35s::E2Fa-DPa, and nac1–1 plants. The results showed that, similar to the 35s::E2Fa-Dpa line, the 4 dpg seedling roots of nac1–1 mutants also showed higher endopolyploidy levels compared with the WT (Figure 5E). These findings indicate that endoreduplication was accelerated by the loss-of-function mutations of NAC1, which affected cell division and thus reduced meristem cells in the nac1 mutant.
In Petunia, the NAC family gene NAM was expressed at the primordial and meristem boundaries, and the mutants failed to develop apical shoots [37]; Arabidopsis cuc1 cuc2 double mutants displayed highly affected shoot apical meristem development and also exhibited fused cotyledons, sepals, and stamens [38,39]. In contrast to the other members of the NAC family, whose expression is restricted to shoot meristem and flower (NAM, CUC2, NAP) or to vascular tissues (such as, CmNACP), as a member of NAC family, NAC1 is highly expressed in roots, and especially, the highest NAC1 expression is restricted to lateral root initiation regions and the root tip (meristem and elongation zone) [32]. Previously data identify NAC1 as a transcription activator in the auxin signaling pathway that regulates genes encoding molecules involved in the specification of lateral root formation [32]. During de novo root organogenesis in Arabidopsis, the NAC1 pathway functioned independently of auxin-mediated explant-specific wounding and root tip emergence [33]. However, the mechanism by which NAC1 regulates root meristem and root growth is unknown. Here, we provide evidence to show that NAC1 maintains root meristem and root growth through directly repressing E2Fa transcription, and loss of NAC1 triggers up-regulation of the E2Fa gene expression, which affects root meristem cell division through the regulation of the endocycle (Figure 6). E2F/DP transcription factors are well-established targets of the universal CYC-CDK-RBR cascade and key regulators of S-phase genes governing cell cycle progression and DNA replication during postembryonic development, especially in root meristem cell divisions [23,40,41]. On the other hand, the E2Fa-DPa complex not only regulates the mitotic cell cycle progression but also plays a role in the endocycle [17,24]. A previous study showed that E2Fa-DPa was a key regulator of the endocycle [17], and overexpression of E2Fa and DPa in Arabidopsis showed a shorter primary root (Figure 5) [24], which is consistent with the elevated E2Fa expression that caused root growth arrest in the nac1–1 mutant (Figure 4). Furthermore, the transgenic lines overexpressing the NAC1 were bigger, with larger leaves, thicker stems, and more lateral roots compared with control plants [32]. Considering that endoreduplication is often associated with cell differentiation such as cell growth and organ enlargement, it is reasonable to speculate that the function of NAC1, as a negative regulator of E2Fa expression, in maintaining the root meristem size and root growth that are achieved through repressing E2Fa-mediated endoreduplication. From all the results presented in this study, we put forward an NAC1-E2Fa signaling pathway that controls root meristem growth. However, there are still a couple of unknown questions. For instance, the hormone auxin plays a pivotal role in establishing the root proximodistal axis including meristem, acting as a local signaling factor [3,4], and the NAC1 gene is induced by auxin [31,32]. Interestingly, the relationship between the NAC1-E2Fa signaling pathway and auxin needs to be clarified in the future. On the other hand, NAC1 acts as a transcription activator in promoting lateral root development [32]. Unexpectedly, we found that NAC1 directly binds to the E2Fa promoter to repress its expression; therefore, it should be interesting to find out if NAC1 might regulate specific gene expression through transcriptional activation or repression that is dependent on its partner factors.
In this study, all wild-type, mutant, and transgenic lines are in the Col-0 background; e2fa [25], E2Fapro:GUS [34], 35s::E2Fa, and 35s::E2Fa-DPa [24], as described previously. Mutant seed stocks used in this study are listed in Supplementary Materials. The nac1–1/e2fa double mutant was obtained by crossing of nac1–1 and e2fa. The nac1–1/nac1–2 F1 double mutant was obtained by crossing of nac1–1 and nac1–2. The nac1–1/E2Fapro:GUS transgenic line was obtained by crossing of nac1–1 and E2Fapro:GUS. Plants were sowed on half-strength Murashige and Skoog medium, and then stratified at 4 °C for 2 days in the dark and then transferred to a phytotron set at 22 °C with a 16 h light/8 h dark photoperiod in vertically oriented Petri dishes. Roots were examined at 5 dpg.
The promoter region and coding sequence of mNAC1 were amplified with Gateway compatible primers (Supplemental Table S1) [30]. The PCR products were first cloned to pEntry vector and then recombined with the binary vector pGWB18 (35s promoter, N-4 × Myc) to generate the 35s::Myc-mNAC1 (mNAC1ox) construct. For the NAC1pro:NAC1 construction, the NAC1 coding region or promoter regions were first cloned to the pEntry vector and then recombined with the binary vector pGWB1 (no promoter, no tag). All the constructs were transformed into Agrobacterium tumefaciens strain GV3101, which was used for transformation of Arabidopsis plants by the floral dip method. Transgenes were selected based on their resistance to hygromycin. Homozygous T3 transgenic plants were used for further experiments. For RT-qPCR analysis, total RNA was extracted form 5 dpg roots using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany), and cDNA was prepared form 1 μg of total RNA with Hiscript III Reverse Transcriptase (Vazyme) [42,43]. The expression levels of target genes were normalized against ACT2. Primers are listed in the Supplementary Materials.
For histochemical β-glucuronidase [44] staining, seedlings were infiltrated with 100 mM sodium phosphate buffer (pH 7.2), 0.1% Triton X-100, 2 mM potassium fericyanide and potassium ferrocyanide, 10 mM EDTA, and 2 mM 5-bromo-4-chloro-3-indolyl-β-giucuronide (X-gluc), and incubated at 37 °C for overnight. Samples were cleared in chloral hydrate and visualized with Olympus BX53 microscopy. The process was performed according to Lv et al. [45].
The full-length coding sequences of NAC1 were amplified with the primers listed in the table in Supplementary Materials and cloned into pGADT7 vector, and the promoter sequences of E2Fa, E2Fb, or E2Fc were cloned into the pAbAi vector. All the constructs used for testing the interactions were transformed into Y1Hgold. The presence of transgenes was confirmed by growth on SD-Ura-Leu plates. Protein interactions were assessed by dropping the yeast transformants on SD-Ura-Leu with Aureobasidin A (AbA) plates. Interactions were observed at 30 °C after 2 days of incubation.
Root meristems were imaged by a Zeiss LSM 900 laser scanning microscope with a 20× objective. For confocal laser scanning microscopy, root meristems were mounted in 10 μg/mL propidium iodide. The process was performed according to the method described by Tian et al. [44]. In addition, to determine the number of cells belonging to the root meristem, root meristematic cortex cells were counted in a file extending from the QC to the first elongated cell excluded [46]. We quantified root cortical cell length in the maturation zone which has root hairs using 20 to 50 cells from 15 to 20 roots for each genetic background with Image J. Image processing was performed with the LSM image-processing software (Zeiss, Jena, Germany). We determined statistical significance by Student’s t test or one-way ANOVA (Tukey’s multiple comparison tests).
The GST-NAC1 (1–199 aa) (pGEX-4T-1) protein was expressed in E. coil BL21 (DE3). We grew BL21 cells at 37 °C in Luria–Bertani (LB) medium in the presence of antibiotics to an OD600 of 0.3 to 0.5. We induced protein accumulation by adding IPTG to a final concentration of 0.3 mM and purified with Glutathione Sepharose 4B (GE Healthcare, Chicago, IL, USA, 17-0756-01) according to the manufacturer’s instructions. EMSA was performed using the LightShift Chemiluminescent EMSA kit (Thermo Scientific, Waltham, MA, USA, 20148) according to the manufacturer’s instructions.
ChIP was performed according to the Anne-Valérie Gendrel et al. [47] protocol. Seedlings at 5 dpg were harvested and crossed-linked with 4% formaldehyde under vacuum infiltration, then halted in 2M Gly. Immunoprecipitated chromatin was analyzed by qPCR. Enrichment was calculated as a ratio of bound sequence over input. The expression levels of target genes were normalized against ACT2. Primers are listed in Table S1.
The NAC1 coding sequences were amplified, and the resulting sequences introduced into pBI221 to place them under the control of the CaMV 35s promoter. The E2Fa, E2Fb, and E2Fc promoter sequences were amplified and introduced into the pGreenII0800-LUC vector. Both recombinant plasmids were then transferred into Arabidopsis protoplasts. The process was performed according to the method described by Yoo et al. [48]. Primers are listed in the Table S1.
For extraction of nuclei, 4 dpg meristem roots were quickly, finely chopped with a sharp razor blade in 1 mL buffer 1 (100 mM citric acid, 0.5% (v/v) polysorbate-20, pH 2–3). The process was performed according to the method described by Li et al. [49]. | true | true | true |
|
PMC9603659 | Yong Wang,Yunxia Guo,Chunhui Duan,Junjie Li,Shoukun Ji,Huihui Yan,Yueqin Liu,Yingjie Zhang | LncGSAR Controls Ovarian Granulosa Cell Steroidogenesis via Sponging MiR-125b to Activate SCAP/SREBP Pathway | 12-10-2022 | lncGSAR,MiR-125b,ceRNA,granulosa cell,steroidogenesis | Long non-coding RNAs (lncRNAs) have been shown to play important roles in livestock fecundity, and many lncRNAs that affect follicular development and reproductive diseases have been identified in the ovary. However, only a few of them have been functionally annotated and mechanistically validated. In this study, we identified a new lncRNA (lncGSAR) and investigated its effects on the proliferation and steroidogenesis of ovine granulosa cells (GCs). High concentrations of glucose (add 33.6 mM glucose) caused high expression of lncGSAR in GCs by regulating its stability, and lncGSAR overexpression promoted GCs proliferation, estrogen secretion, and inhibited progesterone secretion, whereas interference with lncGASR had the opposite effect. Next, we found that the RNA molecules of lncGSAR act on MiR-125b as competitive endogenous RNA (ceRNA), and SREBP-cleavage-activating protein (SCAP) was verified as a target of MiR-125b. LncGASR overexpression increased the expression of SCAP, SREBP, and steroid hormone-related proteins, which can be attenuated by MiR-125b. Our results demonstrated that lncGSAR can act as a ceRNA to activate SCAP/SREBP signaling by sponging MiR-125b to regulate steroid hormone secretion in GCs. These findings provide new insights into the mechanisms of nutrient-regulated follicle development in ewes. | LncGSAR Controls Ovarian Granulosa Cell Steroidogenesis via Sponging MiR-125b to Activate SCAP/SREBP Pathway
Long non-coding RNAs (lncRNAs) have been shown to play important roles in livestock fecundity, and many lncRNAs that affect follicular development and reproductive diseases have been identified in the ovary. However, only a few of them have been functionally annotated and mechanistically validated. In this study, we identified a new lncRNA (lncGSAR) and investigated its effects on the proliferation and steroidogenesis of ovine granulosa cells (GCs). High concentrations of glucose (add 33.6 mM glucose) caused high expression of lncGSAR in GCs by regulating its stability, and lncGSAR overexpression promoted GCs proliferation, estrogen secretion, and inhibited progesterone secretion, whereas interference with lncGASR had the opposite effect. Next, we found that the RNA molecules of lncGSAR act on MiR-125b as competitive endogenous RNA (ceRNA), and SREBP-cleavage-activating protein (SCAP) was verified as a target of MiR-125b. LncGASR overexpression increased the expression of SCAP, SREBP, and steroid hormone-related proteins, which can be attenuated by MiR-125b. Our results demonstrated that lncGSAR can act as a ceRNA to activate SCAP/SREBP signaling by sponging MiR-125b to regulate steroid hormone secretion in GCs. These findings provide new insights into the mechanisms of nutrient-regulated follicle development in ewes.
High reproductive efficiency (i.e., litter size) is the core of the sheep breeding industry, and the productivity of a breeding ewe determines its commercial value. Normal proliferation and steroidogenesis of follicle granulosa cells (GCs) are crucial for ovarian follicular development, oocyte maturation, and subsequent embryonic development [1,2]. Therefore, the discovery of genetic regulatory factors involved in GCs function is of great significance for improving ovulation rates and ewe fertility. Normal glucose metabolism in GCs is essential for oocyte development and maturation as well as the protection of GCs development [3,4]. However, chronically high glucose levels have deleterious effects on the structure and function of the ovary, especially oocytes and GCs during folliculogenesis [5]. These complex cellular and development processes depend on the precise spatiotemporal expression of regulatory factors. Our previous studies found that glucose dose had a direct effect on GCs proliferation, apoptosis, and steroid hormone secretion, and glycolytic metabolites were positively correlated with steroid hormone secretions [6,7]. This confirms the connection between GCs steroidogenesis and glucose availability, but the molecular mechanisms underlying these changes require further study. Long non-coding RNAs (lncRNAs) are a class of non-coding RNAs that are longer than 200 nucleotides in length and no protein-coding potential, which are involved in numerous important biological processes [8]. In reproductive activities, lncRNAs played diverse roles in the regulation of follicular development [9], oocyte maturation [10], GCs differentiation [11], and reproductive diseases [12,13]. For example, NEAT1, a lncRNA, was demonstrated to facilitate ovarian GCs proliferation and inhibit apoptosis by regulating the IGF1 [14]. At the post-transcriptional level, lncRNA can serve as efficient microRNA (miRNA) sponges—termed competing endogenous RNAs (ceRNAs)—that interact with miRNA to regulate gene expression [15]. For example, FDNCR sponges miR-543-3p in GCs and prevents miR-543-3p from binding to the DCN 3′ UTR, resulting in DCN transactivation and TGF-b pathway inhibition and promotion of GCs apoptosis in Hu sheep [16]. In addition, lncRNA NORHA acts as a “sponge” that directly binds to the miR-183-96-182 cluster to induce apoptosis in porcine ovarian GCs [17]. Although functions of these lncRNAs have been partially characterized, most of their roles for steroid hormone secretion are still poorly understood. In this study, to explore the role of lncRNAs in regulating glucose-stimulated ovarian GCs steroidogenesis, ovarian GCs cultured with different glucose doses (optimum glucose concentration (8.4 mM) and high glucose concentration (33.6 mM)) [6,7] were used for RNA sequencing (RNA-seq). Based on this result, we investigated the function of a novel lncRNA named granulosa cell steroidogenesis-associated RNA (GSAR). Moreover, the function and underlying regulatory mechanisms of lncGSAR were further explored, and our results elucidated that mechanism of lncGSAR served as a ceRNA regulating ovarian GCs proliferation and steroid hormone secretion. These results provide some clues to the lncRNAs mechanisms of regulation in nutrient-stimulated follicle development and ovulation.
There are 461 (281 upregulated and 180 downregulated) differentially expressed lncRNAs (Figure 1A, Table S3), 94 (47 upregulated and 47 downregulated) differentially expressed miRNAs (Figure 1B, Table S4), and 796 (379 upregulated and 417 downregulated) differentially expressed genes (Figure 1C, Table S5) that were identified with a p-value < 0.05 as the cut-off. To determine the possible functional significance of observed changes in lncRNA levels in the optimum and high-glucose-induced GCs groups, a gene ontology (GO) term enrichment analysis was performed. We found the differentially expressed lncRNAs were found to be similar and significantly associated with steroid hormone stimulus in biological process term enrichment (Figure S1). Interestingly, the Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of lncRNA target genes also found the ovarian steroidogenesis pathway and steroid hormone biosynthesis pathway were both enriched in comparison group (Figure S2). These findings suggest that steroidogenesis may play critical roles in glucose-induced function of GCs. Protein-coding RNAs (mRNAs) and lncRNAs, which share the common miRNA response elements (MREs), can both compete for binding to miRNAs and regulate each other. In this study, we constructed a putative lncRNA–miRNA–mRNA crosstalk network involved in GCs steroidogenesis. A novel differentially expressed lncRNA (the intergenic lncRNA TCONS_00219351 is located on chromosome chr26: 25174136-25228108, and its RNA sequence is shown in Figure S3) was served as a candidate. The expression of lncRNA TCONS_00219351 was significantly up-regulated in the high-glucose group compared with the optimal-glucose concentration group, and the analysis of qRT-PCR also verified this result (Figure 1D). Thus, we suspected that lncRNA TCONS_00219351 plays an important role in steroid hormone synthesis of GCs and hereinafter refer to this lncRNA as lncGSAR for convenience. We further investigated the subcellular localization of lncGSAR, and the RT-PCR result confirmed that it is an RNA molecule present in the cytoplasm and nucleus (Figure 1E). The lncGSAR-binding miRNA were predicted as candidates using RNAhybrid and miRanda software. A ceRNAs network (lncGSAR–miRNA–mRNA) with 1 miRNA and 10 mRNAs was constructed in GCs (Figure 1F, Table S6). Notably, SREBP-cleavage-activating protein (SCAP) gene from lncRNA–miRNA–gene network was a critical sensor of glutamine, glucose, and sterol levels [18]. SCAP regulates sterol production by activating the sterol regulatory element-binding proteins (SREBPs) signaling pathway, which plays an important role in the subsequent steroid hormone synthesis [19].
To validate the effect of different glucose on the expression of lncGSAR, we treated g GCs with high-concentration glucose with or without actinomycin D (ACTD), which is an RNA transcription inhibitor. As shown in Figure 2A, the lncGSAR level was increased after high-concentration glucose treatment; however, the enhanced lncGSAR fails to be rescued by ACTD, which suggests high-concentration glucose regulated lncGSAR expression dependent on enhancing its stability. To further explore the mechanism of glucose enhances lncGSAR stability, we cloned the full-length lncGSAR sequence after luciferase element of pMIR vector to obtain pMIR-luciferase lncGSAR and cloned the lncGSAR promoter before the luciferase element of Pgl3 vector to obtain Pgl3-lncGSAR promoter luciferase plasmids, respectively (Figure 2B). It is no surprise that the luciferase activity of Pgl3-lncGSAR promoter luciferase cannot be changed, yet the activity of pMIR-luciferase lncGSAR was increased under high-level glucose treatment (Figure 2C,D). Thus, these data indicate expression of lncGSAR was increased through reinforcing its stability under high-glucose stress.
In order to assess the function of lncGSAR in GCs, lncGSAR overexpression vector (pcDNA3.1-lncGSAR) or lncGSAR small interfering RNA (si-lncGSAR) were constructed and transfected into GCs. The cell counting kit-8 (CCK-8) assay showed that overexpression of lncGSAR significantly increased GCs viability (Figure 3A). The 5′-bromo-2′-deoxyuridine (BrdU) detection also demonstrated that the proliferation rate of lncGSAR overexpression cells was significantly increased compared with that of the control (Figure 3C). Conversely, the opposite result was observed by lncGSAR interference (Figure 3B,C), indicating that lncGSAR can facilitate GCs proliferation. To further investigate the role of lncGSAR in secretion of steroid hormones, we measured the concentrations of estradiol (E2) and progesterone (P4) in GCs by enzyme-linked immunosorbent assay (ELISA). The result revealed that overexpression of lncGSAR significantly increased concentrations of E2 (Figure 3D) but decreased concentrations of P4 (Figure 3F). In addition, we found that lncGSAR overexpression significantly upregulated expression levels of steroidogenesis-related (CYP11A1, CYP19A1, and 3β-HSD) and sterol regulation-related (SCAP and SREBP) mRNAs (Figure 3H). Western blot analysis also showed that overexpression of lncGSAR significantly improved the expression levels of steroidogenesis-related proteins and sterol regulatory element proteins (Figure 3J). On the contrary, lncGSAR interference inhibited E2 secretion (Figure 3E) but promoted P4 secretion (Figure 3G) as well as significantly downregulated mRNA and protein expression levels of steroidogenesis-related and sterol regulation-related genes (Figure 3I,J). These results suggest that lncGSAR might play a crucial role in the regulation of steroidogenesis in GCs.
MiR-125b is the core component of lncRNA–miRNA–gene network and has been reported to play a critical role in reproductive hormone biosynthesis and follicle development [20]. In addition, MiR-125b regulated GCs apoptosis in the yak ovary by targeting BMPR1B [21]. In the present study, the lncRNA–miRNA–gene network suggests both the lncGSAR and MiR-125b exhibits the deeply potential to interact with each other (Figure 1F). Furthermore, the RNA-seq showed that the expression of MiR-125b was significantly downregulated in the high-glucose group, and the analysis of qRT-PCR also verified this result (Figure 4A). There was a significant negatively correlation between the MiR-125b expression and lncGSAR expression (Figure 4B). To explore the regulatory relationship between lncGSAR and MiR-125b, granulosa cells were transfected with different doses of MiR-125b mimics (0, 300, 500, and 800 ng) for 48 h. We found that the expression levels of lncGSAR decreased with increasing doses of MiR-125b mimics (Figure 4D), while expression levels of MiR-125b gradually increased (Figure 4C). Next, we added MiR-125b inhibitor to the MiR-125b mimic group and MiR-NC group, respectively, and found that MiR-125b inhibitor could rescue MiR-125b-induced downregulation of lncGSAR expression level (Figure 4E). It is worth noting that addition of MiR-125b mimic reversed the upregulation of lncGSAR expression by high glucose so that the mRNA expression level of lncGSAR had no significant difference between the high-glucose group and the optimal-concentration group (p < 0.05) (Figure 4F). To further investigate the MiR-125b directly binding to lncGSAR, dual-luciferase reporter constructs containing the miRNA response element (MRE; wildtype (WT)) and mutant (MT) plasmid were co-transfected with MiR-125b mimics into GCs (Figure 4G). Luciferase activity of the lncGSAR-WT was dramatically decreased; nevertheless, MiR-125b failed to inhibit the lncGSAR-mut luciferase activity. (Figure 4H). Meanwhile, the RNA–RNA binding assay showed that MiR-125b directly contacts with lncGSAR and not the control RNA GAPDH (Figure 4I).
To further explore the critical functions of MiR-125b in GCs, we transfected GCs with miRNA mimics or miRNA-negative control (MiR-NC) and then monitored the proliferation status of cells using CCK-8 assay, BrdU staining, and flow cytometry analysis. CCK-8 and BrdU staining demonstrated that the proliferation rate of MiR-125b-transfected GCs was significantly reduced compared with that of the MiR-NC-transfected cells (Figure 5A,B). Flow cytometry analysis of the cell cycle revealed that GCs transfected with the MiR-125b mimic could elevate the ratio of cells that progressed to the G1 phase and reduced the ratio of cells that progressed to the S phase (Figure 5C–H). After ELISA, we found MiR-125b overexpression significantly inhibited concentrations of E2 but significantly promoted production of P4 (Figure 5I,J). The expression levels of steroidogenesis- related (CYP11A1, CYP19A1, and 3β-HSD) and sterol regulation-related (SCAP and SREBP) mRNAs were downregulated in MiR-125b overexpression GCs compared to MiR-NC transfected GCs (Figure 5K). In addition, overexpression of MiR-125b reduced the expression level of steroidogenesis-related proteins and sterol regulatory element proteins (Figure 5J). Together, these results suggest that MiR-125b inhibits GC proliferation and steroidogenesis.
Based on the above ceRNA (lncGSAR–MiR-125b–mRNA) network, several genes (e.g., SCAP and GPRC5A) were predicted to be MiR-125b target genes. SCAP, a transmembrane structural protein located on the endoplasmic reticulum and functioning as a sterol sensor, attracts our attention [22]. Thus, a SCAP was selected as a candidate target of the MiR-125b for further study. In the present study, the RNA-seq and qRT-PCR both showed that the expression of SCAP was significantly upregulated in the high-glucose group (Figure 6A). There was a significant negatively correlation between the MiR-125b expression and SCAP expression (Figure 6B). Meanwhile, the proliferation rate and E2 concentrations of MiR-125b transfected GCs were significantly downregulated (Figure 6C,D), while the P4 concentration was upregulated (Figure 6E). However, overexpression of SCAP alleviated the reduced proliferation rate and concentrations of E2 or increased concentrations of P4 driven by MiR-125b (Figure 6C–E). Overexpression of SREBP also reversed MiR-125b-driven inhibition of proliferation rate and E2 concentration or promotion of P4 concentration (Figure 6F–H). In addition, we found MiR-125b overexpression significantly decreased the protein expression level of SCAP and simultaneously downregulated SREBP protein, and the SCAP and SREBP overexpression efficiency were shown in Figure 6I,J. Next, luciferase reporter plasmids containing the SCAP-WT MRE motif or the mutated versions were constructed and co-transfected with MiR-125b mimics into GCs (Figure 6K). MiR-125b significantly reduced luciferase activity of the reporter containing the MRE-WT motif, while SCAP-MT luciferase reporter activity was unchanged under the MiR-125b overexpression (Figure 6L).
We investigated the potential functions of SCAP in GCs. Here, we transfected SCAP overexpression vector (pcDNA3.1-SCAP) or SCAP small interfering RNA (si-lncGSAR) into GCs. The CCK-8 assay and BrdU staining showed that overexpression of SCAP significantly promoted GC proliferation rate compared with that of the control (Figure 7A,C). Conversely, proliferation rate was significantly inhibited after SCAP KD (Figure 7B,C), indicating that SCAP promotes GCs proliferation. To assess whether SCAP affects steroid hormone levels in GCs, we measured the concentrations of E2 and P4 in GCs after transfection pcDNA3.1-SCAP or vector control for 48 h by ELISA. The result revealed that overexpression of SCAP significantly increased concentrations of E2 (Figure 7D) but decreased concentrations of P4 (Figure 7F). In addition, we found that SCAP overexpression significantly upregulated expression levels of steroidogenesis-related (CYP11A1, CYP19A1, and 3β-HSD) and sterol regulation-related (SCAP and SREBP) mRNAs (Figure 7H). Western blot analysis also showed that overexpression of lncGSAR significantly improved the expression levels of steroidogenesis-related proteins and sterol regulatory element-binding proteins (Figure 7J). On the contrary, SCAP interference inhibited E2 secretion (Figure 7E) but promoted P4 secretion (Figure 7G) as well as significantly downregulated mRNA and protein expression levels of steroidogenesis-related and sterol regulation-related genes (Figure 7I,J). Together, these results suggest that SCAP is involved in GCs proliferation and steroidogenesis via activate SREBP.
The established binding mechanism between MiR-125b/lncGSAR and MiR-125b/SCAP inspired us to further determine how lncGSAR regulates the direct targeting of MiR-125b to SCAP and whether lncGSAR regulated steroidogenesis-related proteins’ and sterol regulatory element-binding proteins’ expression is dependent on SCAP. Therefore, we firstly performed Western blot and dual-luciferase assays on GCs after co-transfection of lncGSAR, MiR-125b, and specific siRNA of SCAP. We found that overexpression of lncGSAR could restore the inhibition of steroidogenesis-related proteins’ and sterol regulatory element-binding proteins’ expression exerted by MiR-125b. Meanwhile, lncGSAR regulates these proteins’ expression, which is dependent on its regulation of SCAP (Figure 8A). Additionally, a dual-luciferase reporter analysis similarly demonstrated that lncGSAR blocks the direct binding of MiR-125b and its target SCAP (Figure 8B). To further explore whether lncGSAR affects the binding of MiR-125b to its target SCAP, we labeled MiR-125b and lncGSAR with biotin and digoxigenin, respectively, as well as performed RNA-RNA binding assays. We found that biotin-labeled MiR-125b was associated with SCAP but not unlabeled MiR-125b. Furthermore, digoxigenin-labeled-sense-lncGSAR disrupted the binding between MiR-125b and SCAP but not for the digoxigenin-labeled antisense lncGSAR (Figure 8C). Taken together, all the results support the idea that lncGSAR acts as an MiR-125b sponge to protect the SCAP from the attack of MiR-125b, thereby promoting cell proliferation and E2 secretion and inhibiting P4 secretion.
Ovarian follicular development is a highly coordinated and nutrition-sensitive process [23], and glucose is an important energy substrate for metabolism in the follicles of many species [24,25,26,27]. Previous studies confirm that short-term dietary supplementation of ewes during the luteal phase can increase fertility due to elevated glucose and steroid hormone concentrations in follicles [19,28,29], whereas glucose has deleterious effects on ovary structure and function at higher concentrations [5,30,31]. Our previous study also demonstrated that glucose dose had a significant effect on GCs function, with 8.4 mM representing the optimal glucose concentration for GCs to secrete steroid hormones in vitro and a higher concentration of glucose (33.6 mM) inhibiting GCs proliferation and steroidogenesis [6,7]. Therefore, there is a potential regulatory relationship between glucose metabolism and steroid hormone synthesis, and further investigation is required to validate the molecular mechanism of glucose-induced GCs differentiation and steroid hormone secretion. In the current study, we identified a candidate lncRNA (lncGSAR) involved in ovine follicular steroidogenesis and firstly reported the expression, function, and mechanism of lncRNAs in glucose-induced GCs steroidogenesis. Therefore, this highlighted the importance of lncRNAs in regulating ewe fertility, which may provide new perspectives on the molecular mechanisms by which nutrition regulates reproduction in ewes. With the development of next-generation high-throughput sequencing, multiple lncRNAs have been identified and been shown to play different roles in regulating fecundity. However, studies on lncRNAs related to the regulation of follicle development mostly come from human reproductive diseases (such as polycystic ovary syndrome [13,32] and ovarian cancer [33]), and most of them focus on the effects of lncRNAs on GCs proliferation, apoptosis, and differentiation [34]. Our study demonstrated that lncGSAR was mainly expressed in high-glucose group GCs and controlled follicular development by regulating GCs steroidogenesis. This suggests that lncGSAR is involved in sheep nutrient (glucose) stimulated follicle development and regulates GCs hormone secretion. Studies found that lncRNAs containing MREs act as molecular sponges to effectively inhibit miRNA function [35,36]. For instance, lncRNA MLK7-AS1 regulates proliferation, metastasis, and EMT process in ovarian cancer cells by sponging miR-375 [37]. LncRNA OIP5-AS1 sponging miR-34a to promote ovarian carcinoma cell invasion and migration [38]. Similarly, using bioinformatics, dual-luciferase reporter assay, and RNA-RNA binding assay, we further confirmed that lncGSAR can adsorb MiR-125b in GCs to inhibit follicular function. These studies may provide insight into the mechanisms of lncRNA regulates fecundity in sheep. As we all know, miRNAs directly or indirectly regulate follicular function [39] and hormone secretion [40]. MiR-125b is an important miRNA, and many studies have focused on its biological functions in cell proliferation and differentiation [41] as well as cancer cells [42]. Emerging evidence suggests that in both mouse and human GCs, androgens suppress follicular atresia by enhancing MiR-125b expression, which then targets proapoptotic proteins [43]. On this basis, our study revealed that MiR-125b is involved in the regulation of GCs hormone secretion function and further identified the sterol regulation-related gene SCAP as a novel target of MiR-125b. SCAP is a protein with a sterol-sensing domain (SSD) and seven WD domains. In the presence of cholesterol, this protein binds to sterol regulatory element-binding protein (SREBP) and mediates their transport from the endoplasmic reticulum to the Golgi apparatus. SREBP is then proteolytically cleaved and regulates sterol biosynthesis [44,45,46]. In present study, we also found that the SCAP/SREBP signaling pathway can regulate steroid hormone secretion in GCs by enhancing the expression of steroidogenesis-related (CYP11A1, CYP19A1, and 3β-HSD) genes, which is consistent with the role of lncGSAR and inverse to that MiR-125b. Meanwhile, our results showed that overexpression of MiR-125b reduced the protein levels of SCAP and SREBP. Luciferase reporter gene analysis also confirmed that SCAP is a target gene of MiR-125b. These observations suggest that MiR-125b regulates GCs steroid hormone secretion by targeting SCAP. Furthermore, lncGSAR overexpression increased proteins’ expression of SCAP and SREBP, whereas this effect was eliminated by MiR-125b mimics. The dual-luciferase reporter assay and clips test also confirmed again that lncGSAR acts as an MiR-125b sponge to protect the SCAP from the attack of MiR-125b. This result further demonstrated that binding of lncGSAR to MiR-125b regulates GCs state and function by targeting SCAP. However, there are various models for lncRNAs’ functions, and the regulatory mechanism of lncGSAR in vivo requires further study.
All procedures used in this study were approved by the Animal Care and Use Committee of Hebei Agricultural University (Hebei, P.R. China; permit number 2022100). Fresh ewe ovaries (from thin-tailed Han sheep, ages ranged from 1 to 1.5 years) were collected at the local abattoir (Baoding, Hebei, China) and transported to the laboratory within 3 h in a buffered saline solution supplemented with streptomycin/penicillin mixture (1%) maintained at 37 °C [47]. Small, immature follicles between 1 and 3 mm in diameter were punctured with a disposable syringe, with follicular fluid collected from numbers of ovaries (ovaries number > 50) to negate any individual animal effects. The follicle suspensions were pooled, and GCs were harvested immediately after centrifuging at 1000× g for 10 min. GCs were counted with a hemocytometer (Axio Vert. A1, Zeiss, Oberkochen, Germany). Then, the GCs were seeded in cell culture plates (Thermo Fisher Scientific, Waltham, MA, USA) at a density of 2 × 105/well and cultured in Dulbecco’s Modified Eagle Medium (DMEM/F12, Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, USA) and 1% streptomycin/penicillin mixture in a humidified atmosphere at 37 °C and 5% CO2 for 48 h with the medium changed every 24 h.
When cell confluence reached 80% medium, the original medium is removed. Cells were washed with 1 × PBS, and then, all treatments were cultured in DMEM, which was prepared with no serum, glucose, pyruvate, and phenol red (Solarbio, Beijing, China) for 8 h. Subsequently, the cells were supplemented with 8.4 mM and 33.6 mM of glucose cultured for an additional 24 h. The 8.4 mM represented an optimum glucose concentration for the secretion of steroid hormones by ovine GCs in vitro [6,7], and the 33.6 mM group represents 30 times the physiological concentration of glucose in follicular fluid and was used to detect changes in steroid hormones at super-physiological concentrations [48,49,50]. This culture system was developed so that GC retains hormonally responsive aromatase activity and does not luteinize with time in culture [51,52,53,54]. The GCs were collected for subsequent RNA-seq after the 24 h treatment period.
Ovine GCs subjected to optimum groups (n = 3, add 8.4 mM glucose groups) and high groups (n = 3, add 33.6 mM glucose groups) were used for RNA-seq. The total RNA of each sample was isolated using TRIzol reagent (Invitrogen, Life Technologies, Carlsbad, CA, USA). RNA integrity was assessed using the RNA Nano 6000 Assay Kit of the Agilent Bioanalyzer 2100 System (Agilent Technologies, Santa Rosa, CA, USA). Library preparation and Illumina sequencing analysis were performed as previously described, and putative lncRNAs were screened using unknown transcripts [7]. To investigate interactions between lncRNAs and mRNAs, we constructed a complementary pair network comprising mRNA and lncRNA using Cytoscape 3.6.1(By the National Institute of General Medical Sciences, USA) (https://cytoscape.org/, accessed on 11 September 2018) [55]. Heatmaps were generated by using the R package. The raw sequencing dataset supporting the results of this study was submitted to NCBI BioProject (PRJNA825818) (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE200668, accessed on 12 April 2022).
Total RNA was extracted from cultured cells according to the manufacturer’s instructions and supplied with the TRIzol Reagent (Invitrogen, Life Technologies, Carlsbad, CA, USA); cDNA synthesis for RNA (mRNA and lncRNA) was carried out using the Prime Script RT Reagent Kit with gDNA Eraser (Perfect Real Time) (TaKaRa, Otsu, Japan). Primers were designed and synthesized by GenechemBio (Shanghai, China), and primers used for quantitative real-time PCR are listed in Table S1. Real-time quantitative PCR reactions were performed on a Bio-Rad CFX96 Real-Time Detection System using an iTaq Universal SYBR Green Supermix Kit (Bio-Rad Laboratories Inc., Hercules, CA, USA). Data analyses were performed using the 2ΔΔCt method. β-actin was used as an internal control for mRNA and lncRNA, and U6 was used as an internal control for miRNA.
For nuclear and cytoplasmic RNA separation, 1 × 106 cells were collected and extracted using a Paris kit (Life Technologies, Pleasanton, CA, USA), according to the manufacturer’s instructions. U6 and GAPDH were used as positive controls for the nucleus and cytoplasm, respectively.
The granulosa cells were incubated overnight (at 60–70%, confluence), and transfection or co-transfection was performed using Lipofectamine 2000 (Invitrogen, Shanghai, China) for 48 h. The SCAP overexpression construct was generated by amplifying the SCAP coding sequence, which was subsequently integrated into the HindIII/KpnI restriction sites of the pcDNA3.1 overexpression plasmid (named pcDNA3.1-SCAP). The SREBPs overexpression plasmid was generated by amplifying the SREBPs coding sequence, which was subsequently integrated into the HindIII/KpnI restriction sites of the pcDNA3.1 overexpression plasmid (named pcDNA3.1-SREBPs). The lncRNA GSAR sequence was amplified by PCR and then subcloned into the pCDNA3.1 vector, generating pCDNA3.1-lncGSAR. The empty pcDNA3.1 vector was used as control plasmid. MiR-125b binding sites in SCAP 3′UTR or lncGSAR were amplified by PCR using a cDNA template synthesized from total RNA. Then, the PCR products were subcloned into XhoI/XbaI restriction sites in the pmirGLO dual-luciferase reporter vector to generate the pmirGLO-Luc-SCAP reporter and pmirGLO-Luc-lncGSAR reporter. The MiR-125b mimics, mimic-negative control, siRNA target against the SCAP gene (si-SCAP), siRNA target against the lncGSAR gene (si-lncGSAR), and siRNA nonspecific control were designed and synthesized by RiboBio (Guangzhou, China). The GCs were incubated overnight (at 60–70%, confluence). All transient transfections were performed with Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA, USA), according to manufacturer’s direction. The primers and oligonucleotide sequences used in this study are listed in Tables S1 and S2.
Primary GCs were seeded in 96-well plates and cultured in growth medium. After being transfected, cell proliferation was monitored using a TransDetect CCK (TransGen Biotech, Beijing, China), according to the manufacturer’s protocol. Absorbance was measured using a Model 680 Microplate Reader (Bio-Rad, Hercules, CA, USA) by optical density at a wavelength of 450 nm.
For the flow cytometry analysis of the cell cycle, GCs were seeded in 12-well plates. When the cells grew to a density of 50% confluence, they were transfected with overexpression plasmids. After transfection for 24 h, the cells were collected and fixed overnight in 70% ethanol at 4 °C. Subsequently, the fixed cells were stained with a 50 µg/mL propidium iodide solution (Sigma Life Science, St. Louis, MO, USA) containing 10 µg/mL RNase A (Takara, Japan) and 0.2% (v/v) Triton X-100 (Sigma Life Science, St. Louis, MO, USA) and then incubated in the dark at 37 °C for 30 min. Flow cytometry analysis was performed on a BD Accuri C6 flow cytometer (BD Biosciences, San Jose, CA, USA), and data were processed using the FlowJo7.6 software (Stanford University, Stanford, CA, USA).
BrdU (Sigma, #B5002) incorporation assay was carried out by following the protocol. Briefly, BrdU was diluted to a final concentration of 0.03 mg/mL with fresh DMEM and then applied onto the cells grown on slices. Cells were incubated with 1.5 M HCl followed by 5-min fixation in 70% cold ethanol. Fluorescence staining with anti-BrdU antibody was then conducted. The slices images were captured by a Zeiss Axio Observer confocal microscope.
Granulosa cells were seeded in six-well plates and transfected with overexpression plasmid, siRNA, or miRNA mimics for 48 h, and the medium was collected for subsequent measurements. The E2 concentration of the culture medium was analyzed using a commercial radioimmunoassay (RIA) kit (H102, Nanjing Jiancheng Bioengineering Institute, Nanjing, China). The P4 concentration was also measured using RIA kit (H089, Nanjing Jiancheng Bioengineering Institute, Nanjing, China).
Granulosa cells were seeded in six-well plates and transfected with overexpression plasmid, siRNA, or miRNA mimics for 48 h. Cells were harvested, washed with 1 × phosphate buffered saline (PBS), and lysed in RIPA lysis buffer. Equal amount of proteins were separated at sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). After incubation with the indicated primary and secondary antibodies, signals were visualized by ECL. Membrane was then ready for scanning by Image studio system. Protein quantification was conducted by ImageJ software. The primary antibodies used were anti-CYP11A1 (1:1000; catalogue no. ab67355, Abcam, Cambridge, UK), anti-CYP19A1 (1:1000; catalogue no. ab18995, Abcam, Cambridge, UK), anti-SCAP (1:1000; catalogue no. ab190103, Abcam, Cambridge, UK), anti-SREBP1 (1:1000; catalogue no. ab3259, Abcam, Cambridge, UK), anti-Argonaute-2 (1:1000; catalogue no. ab186733, Abcam, Cambridge, UK), and anti-β-actin (1:10,000; catalogue no. 66009-1-Ig, Proteintech, Chicago, IL, USA). The goat anti-rabbit IgG (H + L)-HRP (1:5000; catalogue no. ab6721, Abcam, Cambridge, UK) were used as a secondary antibody.
Approximately 3 × 104 GCs were plated onto 24-well tissue culture plates 24 h before transfection. Cells were transfected with a mixture of Renilla luciferase and indicated luciferase reporters using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). Forty-eight hours after transfection, the cells were harvested and subjected to an assay by using the Dual Luciferase Reporter Assay system (Promega, Madison, WI, USA). The luciferase activity was detected by Fluorescence/MultiDetection Microplate Reader (BioTek, Winooski, VT, USA). The relative luciferase activities were normalized with the Renilla luciferase activities.
Biotin-labeled RNA was synthesized in vitro using Biotin RNA Labeling Mix (Roche, St Louis, MO, USA, 11685597910). After treatment with RNase-free DNase I, Biotin-labeled RNA was heated at 70 °C for 15 min followed by 2 min incubation on ice to recover the secondary structure of RNA. The RNA was then incubated with streptavidin agarose beads (Invitrogen, Carlsbad, CA, USA) overnight. The next day, the RNA-RNA complexes were pulled down and collected by streptavidin agarose beads. After that, the immunoprecipitated RNA was eluted, isolated, and reverse transcribed to cDNA for the subsequent qRT-PCR analysis.
All experiments were performed at least three times. Data are presented as means ± standard error of the mean based on three independent experiments. All data were normally distributed, and variance was similar between the statistically compared groups. Statistical analyses were performed using SPSS version 22.0 (SPSS Inc., Chicago, IL, USA). Statistical differences were determined by one-way analysis of variance (ANOVA). Tukey’s test was used to estimate the significance of the results. A p-value < 0.05 were considered statistically significant.
In conclusion, we identified a novel lncGSAR that acts as a sponge for MiR-125b to activate the SCAP/SREBP pathway, resulting in promoting granulose cell proliferation and steroidogenesis. Therefore, this study identified a candidate lncRNA (lncGSAR) involved in ovine fecundity, providing insights into the regulatory mechanisms by which glucose regulates follicular development and a basis for new strategies regulating animal reproduction by nutrients. | true | true | true |
|
PMC9603713 | 36073960 | Nancy León-Montes,Jessica Nava-Galeana,Diana Rodríguez-Valverde,Jorge Soria-Bustos,Roberto Rosales-Reyes,Sandra Rivera-Gutiérrez,Hidetada Hirakawa,Miguel A. Ares,Víctor H. Bustamante,Miguel A. De la Cruz | The Two-Component System CpxRA Represses Salmonella Pathogenicity Island 2 by Directly Acting on the ssrAB Regulatory Operon | 08-09-2022 | CpxRA,SPI-2,ssrAB,Salmonella,cpxRA | ABSTRACT The acquisition of Salmonella pathogenicity island 2 (SPI-2) conferred on Salmonella the ability to survive and replicate within host cells. The ssrAB bicistronic operon, located in SPI-2, encodes the SsrAB two-component system (TCS), which is the central positive regulator that induces the expression of SPI-2 genes as well as other genes located outside this island. On the other hand, CpxRA is a two-component system that regulates expression of virulence genes in many bacteria in response to different stimuli that perturb the cell envelope. We previously reported that the CpxRA system represses the expression of SPI-1 and SPI-2 genes under SPI-1-inducing conditions by decreasing the stability of the SPI-1 regulator HilD. Here, we show that under SPI-2-inducing conditions, which mimic the intracellular environment, CpxRA represses the expression of SPI-2 genes by the direct action of phosphorylated CpxR (CpxR-P) on the ssrAB regulatory operon. CpxR-P recognized two sites located proximal and distal from the promoter located upstream of ssrA. Consistently, we found that CpxRA reduces the replication of Salmonella enterica serovar Typhimurium inside murine macrophages. Therefore, our results reveal CpxRA as an additional regulator involved in the intracellular lifestyle of Salmonella , which in turn adds a new layer to the intricate regulatory network controlling the expression of Salmonella virulence genes. IMPORTANCE SPI-2 encodes a type III secretion system (T3SS) that is a hallmark for the species Salmonella enterica , which is essential for the survival and replication within macrophages. Expression of SPI-2 genes is positively controlled by the two-component system SsrAB. Here, we determined a regulatory mechanism involved in controlling the overgrowth of Salmonella inside macrophages. In this mechanism, CpxRA, a two-component system that is activated by extracytoplasmic stress, directly represses expression of the ssrAB regulatory operon; as a consequence, expression of SsrAB target genes is decreased. Our findings reveal a novel mechanism involved in the intracellular lifestyle of Salmonella , which is expected to sense perturbations in the bacterial envelope that Salmonella faces inside host cells, as the synthesis of the T3SS-2 itself. | The Two-Component System CpxRA Represses Salmonella Pathogenicity Island 2 by Directly Acting on the ssrAB Regulatory Operon
The acquisition of Salmonella pathogenicity island 2 (SPI-2) conferred on Salmonella the ability to survive and replicate within host cells. The ssrAB bicistronic operon, located in SPI-2, encodes the SsrAB two-component system (TCS), which is the central positive regulator that induces the expression of SPI-2 genes as well as other genes located outside this island. On the other hand, CpxRA is a two-component system that regulates expression of virulence genes in many bacteria in response to different stimuli that perturb the cell envelope. We previously reported that the CpxRA system represses the expression of SPI-1 and SPI-2 genes under SPI-1-inducing conditions by decreasing the stability of the SPI-1 regulator HilD. Here, we show that under SPI-2-inducing conditions, which mimic the intracellular environment, CpxRA represses the expression of SPI-2 genes by the direct action of phosphorylated CpxR (CpxR-P) on the ssrAB regulatory operon. CpxR-P recognized two sites located proximal and distal from the promoter located upstream of ssrA. Consistently, we found that CpxRA reduces the replication of Salmonella enterica serovar Typhimurium inside murine macrophages. Therefore, our results reveal CpxRA as an additional regulator involved in the intracellular lifestyle of Salmonella, which in turn adds a new layer to the intricate regulatory network controlling the expression of Salmonella virulence genes. IMPORTANCE SPI-2 encodes a type III secretion system (T3SS) that is a hallmark for the species Salmonella enterica, which is essential for the survival and replication within macrophages. Expression of SPI-2 genes is positively controlled by the two-component system SsrAB. Here, we determined a regulatory mechanism involved in controlling the overgrowth of Salmonella inside macrophages. In this mechanism, CpxRA, a two-component system that is activated by extracytoplasmic stress, directly represses expression of the ssrAB regulatory operon; as a consequence, expression of SsrAB target genes is decreased. Our findings reveal a novel mechanism involved in the intracellular lifestyle of Salmonella, which is expected to sense perturbations in the bacterial envelope that Salmonella faces inside host cells, as the synthesis of the T3SS-2 itself.
Salmonella is a common etiological agent of gastrointestinal disease transmitted by food or water (1, 2). The genus Salmonella is composed of two species: Salmonella enterica, which comprises six subspecies, and Salmonella bongori. So far, over 2,600 different serotypes of S. enterica have been described, which can cause severe gastroenteritis and systemic infections in warm-blooded animals, including humans (3, 4). Pathogenesis of Salmonella enterica serovar Typhimurium (S. Typhimurium) is mostly due to its ability to invade and replicate within intestinal epithelial and phagocytic host cells (5). Major virulence factors of Salmonella are type III secretion systems 1 and 2 (T3SS-1 and T3SS-2) and related effector proteins, encoded in Salmonella pathogenicity islands 1 and 2 (SPI-1 and SPI-2), respectively, which were acquired by Salmonella by horizontal gene transfer events (6–8). SPI-1 and SPI-2 also encode different effector proteins, chaperones, and transcriptional regulators that are necessary for the infection by Salmonella; the effector proteins are translocated into the cytoplasm of host cells by the respective T3SS (5, 7–14). The SPI-1 genes are expressed when Salmonella reaches the intestinal lumen, which is necessary for Salmonella invasion of epithelial cells (9–11). In vitro, the SPI-1 genes are expressed in the late exponential/early stationary phase of growth in nutrient-rich media such as LB, which is thought to somehow mimic the intestinal environment (SPI-1-inducing conditions) (15, 16). Once Salmonella is inside host cells, the expression of SPI-1 is repressed, whereas that of SPI-2 is activated (17–20). The SPI-2 genes are essential for the intracellular replication of Salmonella, in both phagocytic and nonphagocytic cells, in a membrane-bound niche termed the Salmonella-containing vacuole (SCV) (21). In vitro, the SPI-2 genes are expressed in the late stationary phase of growth in nutrient-rich media such as LB, as well as in minimal media like N-minimal medium; minimal media containing low concentrations of calcium, magnesium, and phosphate (SPI-2-inducing conditions) resemble the conditions encountered by Salmonella within host cells (15, 19, 20). A myriad of regulators controls the expression of the SPI-1 and SPI-2 genes (1). HilD, encoded in SPI-1, is the apex of a complex regulatory cascade that activates the expression of the SPI-1 genes and other virulence genes located outside this island (22–25). On the other hand, the SsrAB two-component system (TCS) is the central positive regulator for the SPI-2 genes and other functionally related genes located outside SPI-2 (20, 26–28). The SsrAB system is encoded in the ssrAB operon, located in SPI-2; SsrA is the sensor kinase, and SsrB is the response regulator that directly controls the expression of target genes (1, 27). Interestingly, HilD directly induces the expression of the ssrAB operon, and thus the SPI-2 genes, in the late stationary phase of growth in LB (15, 29), whereas SsrB directly represses the expression of hilD and hilA, and thus the SPI-1 genes, when Salmonella is grown under SPI-2-inducing conditions or when it is inside host cells (30–32), which shows a bidirectional transcriptional communication between SPI-1 and SPI-2. In addition, the regulatory proteins SirA/BarA, HilE, and HilD form an incoherent feed-forward loop that controls the growth cost of virulence factor expression by S. Typhimurium (33). Additionally, the expression of ssrAB under SPI-2-inducing conditions is controlled by several other regulators: positively by OmpR, SlyA, and PhoP and negatively by H-NS, YgdT, and Hha (1, 28, 34–38). CpxRA is a TCS that controls expression of virulence genes in different pathogenic bacteria (39). CpxRA is activated by signals that induce stress in the cell envelope, including protein misfolding in the periplasm, defects in peptidoglycan, elevated pH, hyperosmolarity, alterations in inner membrane lipid composition, indole, copper, ethanol, and EDTA (40, 41). CpxA is the histidine kinase that, when activated, autophosphorylates and then transfers its phosphoryl group to the aspartate residue D51 of the response regulator CpxR; in the absence of activating signals, CpxA acts as a phosphatase of phosphorylated CpxR (CpxR-P) (40–42). CpxR-P regulates the expression of target genes encoding different cellular functions or proteins with distinct activities, such as antibiotic resistance, periplasmic protein folding and degrading factors, peptidoglycan metabolic enzymes, inner membrane proteins and regulators (40, 41). CpxR can also be activated independently of CpxA, when bacteria are grown in the presence of excess carbon, such as glucose or pyruvate. This occurs through the AckA-Pta metabolic pathway, which generates acetyl phosphate from acetyl coenzyme A (acetyl-CoA) with the phosphotransacetylase (Pta) and acetate kinase (AckA) enzymes; the phosphoryl group from acetyl phosphate is transferred to CpxR (41, 42). CpxRA regulates the expression of virulence genes in enteropathogenic Escherichia coli (EPEC), uropathogenic E. coli (UPEC), enterotoxigenic E. coli (ETEC), avian-pathogenic E. coli (APEC), Shigella spp., Legionella pneumophila, S. Typhimurium, Yersinia pseudotuberculosis, and Haemophilus ducreyi (43–51). In S. Typhimurium, the absence of CpxA leads to the phosphorylation of CpxR through the AckA-Pta pathway; CpxR-P represses the expression of the SPI-1 and SPI-2 genes in SPI-1-inducing conditions by affecting the stability of HilD (51). A previous study reported the positive regulation of ssrB (the second gene of the ssrAB operon) by CpxR, mediated by a CpxR-binding site located between positions +19 and +51 with respect to a transcriptional start site located upstream of ssrB (52). However, the role of CpxRA TCS on the promoter located upstream of ssrA remains unknown. In this work, we determined that the TCS CpxRA represses the expression of SPI-2 genes under SPI-2-inducing conditions by the direct action of CpxR-P on the ssrAB regulatory operon, specifically on the promoter located upstream ssrA. Two CpxR-binding sites (proximal and distal from the promoter located upstream of ssrA) were required for the regulation/binding of CpxRA on ssrAB. Consistently, our results show that CpxRA reduces S. Typhimurium replication within RAW264.7 macrophages. Our findings further expand the knowledge about the regulatory mechanisms controlling the intracellular lifestyle of Salmonella.
In order to analyze whether CpxRA controls the expression of the SPI-2 genes in SPI-2-inducing conditions, we firstly quantified by RT-qPCR the expression of the ssrAB regulatory operon in the wild-type (WT) S. Typhimurium strain and its derivative ΔcpxR, ΔcpxA, and ΔcpxRA mutants grown in N-minimal medium (N-MM). The expression of both ssrA and ssrB genes increased 4-fold in the ΔcpxR and ΔcpxRA mutants compared to the WT strain (Fig. 1A), suggesting that CpxR negatively regulates expression of the ssrAB operon. In contrast, decreases of 2- and 3-fold in the expression of ssrA and ssrB, respectively, were detected in the ΔcpxA mutant compared to the WT strain (Fig. 1A). Several reports have demonstrated that both null and truncated mutants in the cpxA gene show high levels of CpxR-P, due to the phosphorylation of CpxR by acetyl-phosphate produced by the AckA and Pta enzymes and the absence of the CpxA phosphatase activity on CpxR-P (43–48, 51). Consistently, the expression of genes positively regulated by SsrAB (ssaB, sseA, ssaG, sifA, sseJ, and pipB) was affected in the ΔcpxA, ΔcpxR, and ΔcpxRA mutants similarly to the expression of the ssrA and ssrB genes (Fig. 1B and C): ssaB, sseA, and ssaG are located in SPI-2, whereas pipB, sseJ, and sifA are located outside SPI-2 (14, 53). A cpxRA-cat transcriptional fusion was used for control of expression in N-MM. In the absence of the CpxA sensor kinase, cpxRA expression increased 2-fold compared to that in the WT (Fig. 1D). In contrast, cpxRA transcription was diminished in the ΔcpxR and ΔcpxRA mutants, corroborating the notion that cpxRA is positively autoregulated (54–56). While transcription of cpxRA was not changed in the Δ(ackA-pta) mutant, the expression of cpxRA was downregulated in the ΔcpxA Δ(ackA-pta) double mutant to levels similar to those of the ΔcpxR and ΔcpxRA mutants, showing that CpxR is constitutively phosphorylated by the AckA-Pta enzymes in a ΔcpxA background (Fig. 1D). Overexpression of the outer membrane lipoprotein NlpE causes the activation of the TCS CpxRA (57, 58). Therefore, we evaluated by RT-qPCR the effect of NlpE-mediated activation of CpxRA on the expression of the SPI-2 genes in the WT S. Typhimurium strain and its ΔcpxRA mutant grown in N-MM. As shown in Fig. 2A, overexpression of NlpE decreased 4-fold the expression of both ssrA and ssrB in the WT strain but not in the ΔcpxRA mutant, corroborating that NlpE-mediated activation is CpxRA dependent (Fig. 2A). Similarly, NlpE overexpression also diminished the expression of the genes activated by SsrAB in the WT strain but not in the ΔcpxRA mutant (Fig. 2B and C). Together, these results show that the CpxRA TCS represses the expression of the ssrAB operon and, as a consequence, the expression of SsrAB target genes under SPI-2-inducing conditions.
In order to determine the cis-acting sequences required for the regulation of ssrAB by CpxRA, the expression of a series of cat fusions carrying different segments of the regulatory region of ssrAB (Fig. 3A) was quantified in the WT S. Typhimurium strain and its derivative ΔcpxA mutant grown in N-MM. The ssrAB-cat-302/+478 fusion, containing the most extended regulatory sequence of ssrAB, from position −302 to position +478 with respect to the transcriptional start site, showed decreased expression levels in the ΔcpxA mutant compared to the WT strain (Fig. 3B), confirming the negative regulation of CpxR on ssrAB. A similar expression pattern was also obtained for the ssrAB-cat-208, ssrAB-cat-106, and ssrAB-cat-55 fusions carrying 5′ deletions, as well as for the ssrAB-cat+336 and ssrAB-cat+240 fusions carrying 3′ deletions, their expression was decreased in the ΔcpxA mutant with respect to the WT strain (Fig. 3B and C). In contrast, the expression levels of the ssrAB-cat+119, ssrAB-cat+69, and ssrAB-cat+10 fusions carrying 3′ deletions were not significantly different between the ΔcpxA mutant and the WT strain (Fig. 3C). The different effects of the absence of CpxA on the ssrAB-cat+240 and ssrAB-cat+119 fusions revealed that the sequence spanning the positions +119 to +240 is required for the repression of ssrAB by CpxRA.
To determine whether CpxRA directly regulates ssrAB, electrophoretic mobility shift assays (EMSAs) were performed with purified CpxR and the DNA region of ssrAB (fragment −302/+478) contained in the ssrAB-cat-302/+478 fusion, which showed regulation by CpxRA (Fig. 3B). Binding reactions were performed with phosphorylated (CpxR-P) or with nonphosphorylated CpxR. As expected, CpxR-P, but not CpxR, bound the fragment −302/+478 at concentrations of 1.0 and 2.0 μM (Fig. 4A). Additionally, CpxR-P bound the fragment carrying the upstream region of cpxP (positive control), a gene regulated by CpxRA (59), but it did not bind the fragment carrying the upstream region of the ssaG gene (negative control). Then, we analyzed binding regions of CpxR-P with different segments of the regulatory region of ssrAB. As shown in Fig. 4B, CpxR-P bound to a fragment spanning positions −55 to +240 of ssrAB. In contrast, CpxR-P did not bind fragments spanning positions −302 to −55, +1 to +119, and +240 to +478 of ssrAB (Fig. 4C). These results reveal that CpxR-P acts on the sequence between positions −55 and +240 of ssrAB. In agreement with the results from EMSAs, two putative CpxR-binding sites were manually found on ssrAB (boxes I and II) (Fig. 4D), according to the CpxR-binding consensus sequence [GTAAA(N)4-8GTAAA] reported for E. coli (60, 61). CpxR box I (GAAAAATTATTTATTAAA) and CpxR box II (GCAAACATCTTTAGTAAT) are located between positions −39 and −22 and positions +198 and +215, respectively, with respect to the transcriptional start site of ssrA. Interestingly, fragments containing only one of the putative CpxR-binding sites, fragment −55/+119 (carrying CpxR box I) and fragment +1/+240 (carrying CpxR box II), were not bound by CpxR-P (Fig. 4B), suggesting that both sites are necessary for binding of CpxR-P on ssrAB. In agreement with this conclusion, the expression of ssrAB-cat fusions carrying only the CpxR-I box (ssrAB-cat+119, ssrAB-cat+69, and ssrAB-cat+10) was not repressed by the absence of CpxA (Fig. 3C). Thus, these results support that CpxR-P binds two sites located on the sequence spanning the positions −55 and +240 of ssrAB.
Since our results revealed that CpxRA represses the expression of SPI-2 genes under conditions resembling the intracellular environment, we hypothesize that CpxRA negatively controls the intracellular replication of Salmonella. To investigate this, we analyzed the replication within RAW264.7 murine macrophages of the WT S. Typhimurium strain and its ΔcpxR, ΔcpxA, and ΔcpxRA mutants. As shown in Fig. 5, the intracellular replication of the ΔcpxR and ΔcpxRA mutants was ~25% higher, whereas the intracellular replication of the ΔcpxA mutant was ~63% lower, than that of the WT strain. Therefore, our results show that CpxR decreases the replication of S. Typhimurium inside macrophages, which can be explained, at least in part, by its negative control of the expression of the ssrAB regulatory operon.
In this study, we found that in S. Typhimurium, the TCS CpxRA represses the expression of SPI-2 and functionally related genes in conditions resembling the intracellular environment (SPI-2-inducing conditions). This regulation is mediated directly on the ssrAB operon, encoding the TCS SsrAB, the central positive regulator for the SPI-2 and functionally related genes. Our results support that CpxR-P binds two sites located on the sequence spanning positions −55 to +240 of ssrAB, where two putative CpxR-binding sites were identified, at positions −39 to −22 (GAAAA-N8-TTAAA) and +198 to +215 (GCAAA-N8-GTAAT). These sequences did not show a high identity with the GTAAA(N)4-8GTAAA tandem repeat described for E. coli (60, 61). However, a perfect match of the consensus CpxR-binding-site on the DNA is not required, because it does not correlate with the strength of the transcription regulation (56). We determined that in the absence of the sequence between positions +119 and +240, carrying one of the identified CpxR-binding sites, the negative regulation of the ssrAB operon by CpxR is lost. Consistently, our results support that CpxR-P requires the two identified sites for binding on ssrAB. Interestingly, repression of ssrAB by the histone-like H-NS protein also requires the two sequences that mediates the repression of this operon by CpxR-P (29). Subramaniam et al. reported that CpxR positively regulated ssrB (the second gene of the ssrAB operon) in the no-carbon-E (NCE) medium, finding a CpxR-binding site (GTAAA-N5-GGAAA) located in the sequence between positions +19 and +51 with respect to a transcriptional start site located upstream of ssrB (52). It is important to note that this CpxR-binding site on ssrB was not tested in our study. Therefore, CpxR seems to regulate the expression of ssrAB by acting on multiple binding sites. The TCS CpxRA has been shown to repress virulence genes in other bacteria, such as Y. pseudotuberculosis, Shigella sonnei, S. Typhimurium, and different pathotypes of E. coli: enteropathogenic, enterohemorrhagic, and avian pathogenic (43–51, 62). CpxR seems to follow two different mechanisms to repress gene expression. (i) CpxR-P recognizes one site on the regulatory sequence overlapping the −35 and −10 boxes and subsequently blocks binding of RNA polymerase to the promoter. This regulation mode has been reported for the csgBAC, csgDEFG, motAB, cheAW, rpoE, tsr, ung, aer, and rovA genes (60, 63–65). (ii) CpxR-P binds to two or more sites on the regulatory region, competing with activators by binding to sites on the promoter and also blocking binding of RNA polymerase. This mechanism has been described for ompF and papBA genes (66, 67). The repression model of CpxR-P on ssrAB would imply competition between CpxR and OmpR, similar to the observed for the porin gene ompF, which is supported by the overlapping of two DNA-binding sites of both regulators on the ssrAB promoter: OmpR, −83/+6 and +130/+146 (35), and CpxR, −55/+1 and +119/+240. To control gene expression, CpxRA respond to signals that induce extracytoplasmic stress, such as protein misfolding in the periplasm, defects in peptidoglycan, elevated pH, hyperosmolarity, alterations in inner membrane lipid composition, indole, copper, ethanol, and EDTA (40, 41). Furthermore, CpxRA seems to act as a switch-off system for the biogenesis of diverse structures localized between the inner and outer membranes, like the type III secretion systems present in many bacteria, such as S. sonnei, Y. pseudotuberculosis, enteropathogenic E. coli, enterohemorrhagic E. coli, and S. Typhimurium (T3SS-1 and T3SS-2) (44–47, 50, 51). Thus, the CpxRA system could be activated by misfolded structural proteins of T3SS-2 when S. Typhimurium resides within a late SCV. Hha and YdgT have been described as negative regulators of SPI-2. These proteins do not bind to DNA; instead, they act as modulators of the transcription by direct interaction with H-NS (68, 69), which is the main repressor of SPI-2 under SPI-2-inducing conditions (27, 70, 71). EIIANtr, a component of the nitrogen-metabolic phosphotransferase system, also acts as a negative regulator of SPI-2; it interacts with SsrB and thus blocks DNA binding of SsrB on target genes (72). Therefore, CpxR is a repressor that directly interacts with the ssrAB regulatory region downregulating the promoter located upstream ssrA. We found that CpxRA inhibits the replication of S. Typhimurium inside mouse macrophages. Interestingly, a previous study showed that the TCS PhoP/PhoQ prevents the overgrowth of S. Typhimurium inside nonphagocytic cells (73). Therefore, it is reasonable to think that CpxRA and PhoP/PhoQ, and possibly other regulators, such as H-NS (70, 71), EIIANtr (72), Hha (68), and YdgT (74), coordinate to restrict intracellular overgrowth of Salmonella within host cells. Our findings further expand the knowledge about the mechanisms mediating the intracellular lifestyle of Salmonella.
Bacterial strains used in this study are listed in Table 1. Bacterial cultures were grown at 37°C in lysogeny broth (LB) at pH 7.5 or in N-minimal medium (N-MM) [5 mM KCl, 7.5 mM (NH4)2SO4, 0.5 mM K2SO4, 1 mM KH2PO4, 100 mM Tris-HCl, 10 μM MgCl2, 0.5% glycerol, and 0.1% Casamino Acids] at pH 7.5. Flasks of 250 mL containing 30 mL of N-MM were inoculated with bacterial suspensions prepared from overnight cultures in LB, adjusted to an optical density at 600 nm (OD600) of 0.05, and then incubated at 37°C in a shaking incubator at 200 rpm to an OD600 of 0.8 to 1.0. When necessary, media were supplemented with ampicillin (200 μg/mL), chloramphenicol (30 μg/mL), or kanamycin (50 μg/mL).
Total RNA was extracted from bacteria grown in N-MM using the hot-phenol method (75). The RNA was pelleted by centrifugation at 19,000 × g for 10 min at 4°C. Pellets were washed three times with cold 70% ethanol and centrifuged at 16,000 × g for 10 min at 4°C. The ethanol was removed, the pellets were air dried for 15 to 20 min in a centrifugal vacuum concentrator (5301; Eppendorf). The RNA was resuspended in diethyl pyrocarbonate (DEPC)-treated water, DNA was removed with Turbo DNA-free (Ambion), and the quality of RNA was assessed using a NanoDrop One instrument (Thermo Scientific) and with a bleach denaturing 2% agarose gel, as previously described (76). cDNA was synthesized using 1 μg of RNA, 5 pmol/μL of random hexamer primers, and 200 U/μL of RevertAid Moloney murine leukemia virus (M-MuLV) reverse transcriptase (RT) (Thermo Scientific). Specific primers were designed with Primer3Plus software and are listed in Table 2. Quantitative RT-PCR was performed in a LightCycler 480 instrument (Roche). The absence of contaminating DNA was tested by the lack of amplification products after 45 qPCR cycles using RNA as the template. Control reactions with no RNA template and with no reverse transcriptase enzyme were run in all experiments. The relative gene expression was calculated using the 2−ΔΔCT method (77). These experiments were performed in triplicate three independent times.
Chloramphenicol acetyltransferase (CAT) assays were performed as described previously (78).
The E. coli strain BL21 transformed with the pQE80cpxR plasmid (79) was used to express and purify His6-CpxR. A flask containing 250 mL of LB with ampicillin (100 μg/mL) was inoculated 1:100 with an overnight culture of E. coli BL21+pQE80cpxR and then incubated at 37°C with shaking to an OD600 of 0.6 to 0.8 (~4 h). Subsequently, isopropyl-β-d-thiogalactopyranoside (IPTG) was added to a final concentration of 1 mM, and the bacterial culture was grown for an additional 3 h at 37°C under the same conditions. Cells were then pelleted by centrifugation and resuspended in 1× phosphate-buffered saline (PBS)–8 M urea, pH 8.0, and lysed by sonication. The lysate was centrifuged, and the soluble fraction was loaded in a nickel-nitrilotriacetic acid (Ni-NTA) agarose column (Qiagen) pre-equilibrated with lysis buffer. After 10 washes with buffer containing 50 mM imidazole (200 mL), His6-CpxR was eluted with 500 mM imidazole (10 mL). Purified protein was dialyzed over 2 h using a cellulose membrane (dialysis tubing, cellulose membrane; Sigma) and a buffer composed of Tris-HCl (20 mM) (pH 7.5), KCl (50 mM), dithiothreitol (1 mM), and 10% glycerol. Then, the His6-CpxR protein was analyzed by SDS-PAGE and Coomassie blue staining, and its concentration was determined by the Bradford procedure; it was stored at −70°C.
EMSAs were performed with purified His6-CpxR and the ssrAB fragments, which were amplified by PCR with primer pairs showed in Table 2. A DNA fragment containing the regulatory region of ssaG, used as negative control, was amplified by PCR using the primer pair ssaG-5′/ssaG-3′ and the pssaG1-cat plasmid as the template. A DNA fragment carrying the cpxP regulatory region, used as a positive control (59), was amplified by PCR using the primer pair cpxP-5′/cpxP-3′ and chromosomal DNA of E. coli MC4100 as the template. PCR products were purified using the QIAquick PCR purification kit (Qiagen). Purified His6-CpxR protein was phosphorylated with 50 mM acetyl phosphate (Sigma-Aldrich) in buffer containing 10 mM magnesium chloride for 1 h at 30°C. The PCR products were incubated with the CpxR-P protein (0 to 2 μM) in a binding buffer containing Tris-HCl pH 7.5 (20 mM), KCl (50 mM), dithiothreitol (1 mM), and 5% glycerol, for 20 min at room temperature, and then were electrophoretically separated in 6% nondenaturing polyacrylamide gels in 0.5% Tris-borate-EDTA buffer at 4°C. The DNA bands were stained with ethidium bromide and visualized under UV light.
RAW264.7 (ATCC TIB-71) mouse macrophages were seeded at a density of 106 cells per well in 24-well tissue culture plates for 24 h. Bacteria were obtained from cultures in N-MM and were opsonized with normal mouse serum in RPMI containing 10% fetal bovine serum (FBS) for 30 min on ice. Bacteria were added to cells at a multiplicity of infection (MOI) of 100. Plates were centrifuged at 5,000 × g for 5 min at 4°C and incubated for 30 min at 37°C under a humidified 5% CO2 atmosphere. Cells were washed three times with 1× PBS and then were incubated in RPMI containing 100 μg/mL gentamicin and 10% FBS over 1 h to eliminate extracellular bacteria. After this time, the gentamicin concentration was decreased to 10 μg/mL, and cells were incubated for an additional 1 h. Infected macrophages were incubated at 37°C for 2 h and 18 h in a humidified atmosphere with 5% CO2. For enumeration of intracellular bacteria, macrophages were washed three times with 1× PBS and lysed with 0.1% Triton X-100 for 15 min. Then, 10-fold serial dilutions were plated onto LB agar plates, which were incubated overnight at 37°C. CFU were counted in the plates. The experiment was performed three times in triplicate. The replication index was obtained by dividing the number of CFU per milliliter at 18 h by the number at 2 h.
All data are means from three independent experiments. Statistical analysis was performed using Prism 8.0 software (GraphPad, Inc., San Diego, CA, USA). One-way analysis of variance (ANOVA) followed by Tukey’s multiple-comparison test and unpaired Student's t test was performed. P values of ≤0.05 were considered statistically significant. | true | true | true |
PMC9603892 | 36000914 | Jie Xu,Chen Mei,Yan Zhi,Zhi-xuan Liang,Xue Zhang,Hong-jun Wang | Comparative Genomics Analysis and Outer Membrane Vesicle-Mediated Horizontal Antibiotic-Resistance Gene Transfer in Avibacterium paragallinarum | 24-08-2022 | A. paragallinarum,outer membrane vesicles,whole genome,antibiotic resistance gene,horizontal gene transfer | ABSTRACT Avibacterium paragallinarum is the etiological agent of infectious coryza, an acute respiratory disease of chickens that is globally distributed and causes serious economic losses for chicken production. A. paragallinarum is a Gram-negative bacterium that releases outer membrane vesicles (OMVs). In this study, a comparative genomic analysis of A. paragallinarum isolate P4chr1 and its OMVs was carried out, and the ability to transfer antibiotic resistance genes (ARGs) via the OMVs was studied. Sequencing and data analyses demonstrated that the genomic size of A. paragallinarum P4chr1 was approximately 2.77 Mb with a 25 kb tolerance island that covered six types of antibiotics and 11 ARGs. The genomic size of its OMVs was approximately 2.69 Mb, covering 97% of the genomic length and almost all the gene sequences of P4chr1. Purified and DNase-treated A. paragallinarum P4chr1 OMVs were cocultured with the antibiotic-sensitive A. paragallinarum Modesto strain on an antibiotic (chloramphenicol, erythromycin, tetracycline, or streptomycin)-containing plate, and the corresponding ARGs were detected in the colonies grown on the plates. However, using an antimicrobial susceptibility test, we found that ARGs delivered by OMVs were not persistent but only appeared transiently on the antibiotic-containing plates. Antibiotic resistance and ARGs were lost by the second bacterial passage. IMPORTANCE The functions and roles of OMVs on ARG and virulent gene transfer and dissemination have been reported in numerous Gram-negative bacteria. However, the role of OMVs in mediating antibiotic resistance in A. paragallinarum has not been reported. This study is the first report to compare the genomic characteristics of OMVs with its parent A. paragallinarum strain and to study A. paragallinarum ARG transfer via OMVs. This work has provided useful data for further studies focusing on nonplasmid ARG transfer mediated by A. paragallinarum OMVs. | Comparative Genomics Analysis and Outer Membrane Vesicle-Mediated Horizontal Antibiotic-Resistance Gene Transfer in Avibacterium paragallinarum
Avibacterium paragallinarum is the etiological agent of infectious coryza, an acute respiratory disease of chickens that is globally distributed and causes serious economic losses for chicken production. A. paragallinarum is a Gram-negative bacterium that releases outer membrane vesicles (OMVs). In this study, a comparative genomic analysis of A. paragallinarum isolate P4chr1 and its OMVs was carried out, and the ability to transfer antibiotic resistance genes (ARGs) via the OMVs was studied. Sequencing and data analyses demonstrated that the genomic size of A. paragallinarum P4chr1 was approximately 2.77 Mb with a 25 kb tolerance island that covered six types of antibiotics and 11 ARGs. The genomic size of its OMVs was approximately 2.69 Mb, covering 97% of the genomic length and almost all the gene sequences of P4chr1. Purified and DNase-treated A. paragallinarum P4chr1 OMVs were cocultured with the antibiotic-sensitive A. paragallinarum Modesto strain on an antibiotic (chloramphenicol, erythromycin, tetracycline, or streptomycin)-containing plate, and the corresponding ARGs were detected in the colonies grown on the plates. However, using an antimicrobial susceptibility test, we found that ARGs delivered by OMVs were not persistent but only appeared transiently on the antibiotic-containing plates. Antibiotic resistance and ARGs were lost by the second bacterial passage. IMPORTANCE The functions and roles of OMVs on ARG and virulent gene transfer and dissemination have been reported in numerous Gram-negative bacteria. However, the role of OMVs in mediating antibiotic resistance in A. paragallinarum has not been reported. This study is the first report to compare the genomic characteristics of OMVs with its parent A. paragallinarum strain and to study A. paragallinarum ARG transfer via OMVs. This work has provided useful data for further studies focusing on nonplasmid ARG transfer mediated by A. paragallinarum OMVs.
Infectious coryza is an acute upper respiratory disease of chickens caused by Avibacterium paragallinarum, a Gram-negative bacterium of the genus Avibacterium within the family Pasteurellaceae. The disease occurs worldwide and leads to serious economic losses in the chicken industry due to the retarded growth of broilers and reduced egg production in layers (1, 2). Disease prevention has involved the use of inactivated multivalent vaccines (based on local prevalent A. paragallinarum serotypes), and selected antibiotics have only been administered to diseased flocks (1, 3, 4). Several studies have reported increased infectious coryza outbreaks and have examined antibiotic resistance profiles (3–5), as well as detected antibiotic resistance genes (ARGs) in A. paragallinarum (3, 5). The increase in multidrug-resistant bacterial pathogens is a global threat to both public (6) and animal health (7, 8). However, the rate of development of drug-resistant bacteria is currently surpassing the rate of development of new antibiotics (9). Furthermore, it is becoming evident that bacterial drug resistance can be transferred from food-producing animals to humans through accumulated drug residue in meat and egg products (7, 8). Thus, a better understanding of the mechanisms associated with drug resistance is critical. Infections caused by Gram-negative bacteria are more difficult to treat than those caused by Gram-positive bacteria, due to the two layers of complex cell membranes that make up the Gram-negative cell wall (10, 11). In order to adapt to adverse environmental conditions, Gram-negative bacteria have evolved globular bi-layered vesicles (diameter, 50 to 500 nm), known as outer membrane vesicles (OMVs), which are produced through the blebbing and pinching-off the bacterial outer membrane without destroying it (6, 12). OMVs contain many components found in the outer membrane of the bacterial cell, such as lipopolysaccharides, phospholipids, membrane proteins, and peptidoglycan components (6, 12). The lumen of the vesicles contains periplasmic proteins, cytosolic components, and nucleic acids (13). OMVs also carry DNA and RNA on their surface, which can be removed by treating OMVs with DNase and RNase, whereas luminal DNA and RNA are not affected by this treatment (14). Recent studies have revealed a novel mechanism by which antibiotic-susceptible bacteria obtain ARGs from ARG donor bacteria (in the same or different species) using OMVs as vehicles, rather than the three traditional routes, namely, natural transformation, transduction, or conjugation by bacterial cells (6). Although these known mechanisms contribute to the gene flow within bacteria, they have restrictions such as limited genetic load, host specificity, and the type of genetic material that is transferred (6). To date, the functions and roles of OMVs on ARG and virulent gene transfer and dissemination have been reported in numerous Gram-negative bacteria, such as Acinetobacter baumannii, Escherichia coli, Porphyromonas gingivalis, and Pseudomonas aeruginosa (15–19). However, the role of OMVs in mediating antibiotic resistance in A. paragallinarum has not been reported. Recently, we sequenced the whole genome of a newly isolated A. paragallinarum strain, P4chr1 (GenBank accession number CP081939), and identified several ARGs in it. In the current study, we aimed at performing comparative genomics analysis between a multidrug-resistant strain of A. paragallinarum P4chr1 and its OMVs, based on genomic sequencing data. In addition, we demonstrated that A. paragallinarum P4chr1 OMVs mediated the transfer of aminoglycoside antibiotic genes to a drug-susceptible strain by horizontal gene transfer (HGT).
The P4chr1 bacterial solution and extracted OMVs were observed by transmission electron microscopy (TEM). Spherical structures were observed around the A. paragallinarum P4chr1 isolates (Fig. 1a), and extracted OMVs showed a similar spherical structure (Fig. 1b). Thus, our findings demonstrated that P4chr1 can secrete OMVs into the environment during growth. The particle size of the OMVs was between 30 and 100 nm, with an average particle size of 40 nm.
The assembled whole-genome sequence revealed that A. paragallinarum P4chr1 harbored a circular chromosomal DNA (2,774,989 bp) with a 41.01% GC content, and it did not carry a plasmid. In total, 2,778 protein-encoding genes were predicted, with a coding percentage of 95.92%. The average gene length was 852 bp. The software ARAGORN was used to predict tRNAs, and the predicted number was 59. The software RNAmmer was used to predict rRNAs, and the predicted number was 15. In addition, the genes were searched against the KEGG, EggNOG, nonredundant (Nr), nucleotide (Nt) and Swiss-Prot databases to annotate the gene description. Among the 2,778 genes predicted in A. paragallinarum P4chr1, 2,778 genes were annotated into the COG database, accounting for 92.22% of the predicted genes, which could be divided into 21 categories. In addition, 1,851 genes were annotated into the KEGG database, accounting for 66.63% of the predicted genes. These genes were divided into 38 metabolic pathway types (Fig. 2a) and are described in detail in Table 1.
The ResFinder data demonstrated that the chromosomal DNA of A. paragallinarum P4chr1 contained 11 ARGs corresponding to 6 categories of antibiotics (the aminoglycoside resistance genes aph6id, aph3ia, aac3iia, ant2ia, and aph33ib, the beta-lactam resistance gene bl2d_oxa1, the MLS [macrolide, lincosamide, and streptogramin B] resistance gene ermT the phenicol resistance genes catP and cmL_e3, the sulfonamide resistance gene sul2, and the tetracycline resistance gene tetB). The ARGs were all concentrated in a 25-kb fragment of the genome. Sequence comparison analysis revealed that the resistance region of A. paragallinarum P4chr1 exhibited high homology to the corresponding region in the chromosomal DNA of Pasteurella multocida strain FCf83 (accession number CP038875) from China (Fig. 3). Antimicrobial susceptibility testing showed that A. paragallinarum P4chr1 was resistant to chloramphenicol, erythromycin, gentamicin, tetracycline, streptomycin, and ampicillin, whereas the recipient strain A. paragallinarum Modesto was sensitive to these antibiotics.
The genomic sequence of the OMVs of A. paragallinarum strain P4chr1 were composed of 162 contigs for 2,691,804 bp with a 40.92% GC content. The base pair numbers in the OMVs were 97.00% of that found in A. paragallinarum P4chr1. The largest contig was 149,543 bp, and the smallest contig was 261 bp. In total, 2,568 protein-encoding genes were predicted. The average gene length was 859 bp. Fig. 2b shows the genomes of the OMVs annotated in various databases. Comparative genomic circle graphs of A. paragallinarum P4chr1 and its OMVs indicated that the similarity of the two genomes was greater than 90% (Fig. 4a). The collinearity results indicated that most of the genomic segments of the OMVs had counterparts in the A. paragallinarum P4chr1 genome (Fig. 4b). Orthologous cluster analysis of A. paragallinarum P4chr1 and its OMVs showed that P4chr1 had 2,546 clusters, while the OMVs had 2,544 homologous clusters, 2,541 of which were shared by P4chr1 and OMVs (Fig. 5). These results indicated that the genome of the OMVs was derived from A. paragallinarum P4chr1. Furthermore, our data demonstrated that the OMVs had almost complete genomic sequences of A. paragallinarum P4chr1, including some virulence genes and ARGs.
First, purified OMVs isolated from A. paragallinarum P4chr1 were inoculated with tryptic soy agar (TSA) and broth (TSB) containing supplements. No bacterial growth was detected after 24 or 48 h of incubation, indicating that the OMVs were free of bacterial contamination. Next, A. paragallinarum Modesto was transformed with the purified OMVs (Table 2). The transformed colonies produced from the resistant plates were tested for ARGs by PCR. A representative agarose gel showing the corresponding ARG bands is shown in Fig. 6. The sequencing results of the PCR products were also consistent with the ARG sequences annotated at NCBI. Antimicrobial susceptibility testing was performed using the donor strain P4chr1, susceptible strain Modesto, and four ARG-transformed Modesto strains. The MIC values in these four strains did not increase compared with the antibiotic-sensitive Modesto strain (Table 3). A. paragallinarum serovars A, B, and C specific antisera were used in the hemagglutination-hemagglutination inhibition (HA-HI) test to serotype the donor strain P4chr1 (serovar A), recipient strain Modesto (serovar C), and ARG-transformed strains. The HI data showed that A. paragallinarum Modesto and the transformed colonies had titers similar to those of the serovar C Modesto antiserum and no reaction with the serovar A 0083 and B 0222 antisera. These findings were confirmed using the PCR-restriction fragment length polymorphism (RFLP) technique. We found that a 1.6-kb fragment in the hypervariable region of Hmtp210 was amplified for P4chr1, Modesto, and transformed colonies. After digestion with the restriction enzyme Bgl II, the PCR products were divided into two bands, 768 and 868 bp for serovar A and 1,284 and 339 bp in the case of serovar C. We found that the donor bacterium P4chr1 was type A and that both the recipient bacterium Modesto and the transformed colonies were type C (Fig. 7).
Both pathogenic and nonpathogenic Gram-negative bacteria secrete vesicles (11), which contain DNA (plasmid, chromosomal, and/or phage-associated) (11, 20, 21). However, it remains unclear whether OMVs contain a complete genome similar to their parental cells or whether OMVs contain all the genetic information of the bacterial genome. Sequencing purified A. paragallinarum OMVs has led to the identification of genomic fragments in OMVs. In this study, we sequenced the multidrug-resistant A. paragallinarum strain P4chr1 and its OMVs and performed a comparative genomics analysis between the two genomic sequences (2.69 Mb and 2.77 Mb). We found that the GC content, number of coding genes, and metabolic pathways of the two genomes were very similar. Indeed, magnification of the comparative genomic circle map was required to see the differences. Collinearity analysis revealed that each segment of the OMV genomic sequence could be found in the corresponding region of the P4chr1 genomic sequence. Analysis of whole-genome orthologous clusters has been an important step in comparative genomics research. Identifying clusters between orthologous clusters and constructing networks help explain the functions and evolutionary relationships of proteins across multiple species (22). Here, the whole-genome orthologous gene cluster analysis revealed that the OMVs and P4chr1 had 2,541 homologous gene clusters, with only a few differences, which may be caused by sequencing errors or fragmentation during OMV genome sequencing. Our findings indicated that the genomic sequence of the OMVs was derived from A. paragallinarum P4chr1 and that the OMVs had almost complete genomic sequences of A. paragallinarum P4chr1, including virulence genes and ARGs. Within the P4chr1 genomic sequence, 11 ARGs were found to be focused in a 25-kb region, forming a structure similar to a tolerance island. BLAST comparison analysis indicated that this sequence was very similar (99%) to a sequence of a field strain of Pasteurella multocida FCf83, isolated from duck in Fujian, China, in December 2015. Furthermore, the gene-coding direction was also the same. These two bacteria have a close genetic affiliation with the same family, Pasteurellaece, thereby allowing for easier gene exchange. A. paragallinarum and P. multocida are important respiratory pathogens in poultry farms, and they can often be isolated in clinical samples at the same time, which may imply the possibility of horizontal transmission of drug-resistant genes between them (23–25). In addition, the two sequences have the same DNA primase and recombinase before and after, and the same integrase, transposase, recombinase, and endonuclease inside the sequence. Presently, we cannot conclusively determine whether the sequence was transferred by transposition or insertion or possibly through the formation of a cyclized structure from the chromosome. In recent years, several important functions of OMVs have been reported, including the intra- and interspecies horizontal transfer of ARGs (8). In 2011, Rumbo et al. reported the horizontal transfer of plasmids carrying the carbapenemase resistance gene OXA-24 in OMVs to Acinetobacter baumannii (15). In 2015, Ho et al. demonstrated HGT mediated by Porphyromonas gingivalis OMVs. This bacterium (a fimA mutant) carried a 2.1-kb ermF-ermAM cassette in its fimA gene, encoding an erythromycin-resistant gene. The cassette was transferred to the fimA gene of another P. gingivalis strain lacking this gene (erm gene) via OMVs isolated from the donor strain (18). Finally, in 2019, Fulsundar et al. proposed an optimized and detailed plan to test and confirm that OMV-mediated ARGs can be transferred to Acinetobacter baumannii without plasmids in OMVs (14). Here, we examined the potential OMV-mediated HGT from the A. paragallinarum P4chr1 strain, which contains ARGs to the antibiotic-sensitive A. paragallinarum strain Modesto. We found that the antibiotic-sensitive Modesto strain successively survived on antibiotic-treated agar plates with ARGs. Furthermore, its transformed colonies passed several verification tests. ARG-PCR analysis demonstrated that the transformed colonies amplified corresponding ARG products, while the HA-HI and PCR-RFLP assays confirmed that the transformed colonies were derived from the recipient cells and not from donor cells (20, 26). The PCR-RFLP test has some limitations, and our previous studies have discussed the availability of this method (27). In this experiment, PCR-RFLP was suitable for distinguishing serovar A strain P4chr1 from serovar C strain Modesto, and agreed with the result from the HA-HI test. However, the overall ARG transformation efficiency mediated by A. paragallinarum OMVs was low compared to that found in previous studies, which reported that transferred ARGs were generally carried in the plasmid (15, 16, 19). Our HGT assays revealed that the highest transformation frequency was 5.42 × 10−7 (Table 2) in the streptomycin group and that there was a higher copy number of streptomycin-resistant genes in the donor strain P4chr1. Other studies have suggested that OMVs contain only partial genomic fragments, such that some fragments do not contain intact ARGs, and not every vesicle contains DNA (28). Tran and Boedicker hypothesized that the ability to acquire DNA may depend on the species of the donor/recipient bacteria (29). In our antimicrobial susceptibility test, the MIC values produced by the transformed colony strains did not increase compared with the antibiotic-susceptible strain Modesto, suggesting that the ARGs transferred by donor OMVs are not persistent in recipient cells after passaging. This unexpected result exposed a poor passaging ability of the transformed colonies. It has been proposed that once a gene has been transferred into its recipient, it must be integrated into the chromosomal DNA in order to persist within the cells (18). Ho et al., for example, demonstrated that the erm gene in the vesicles of the fimA mutant was flanked with fimA sequences at both ends and that homologous DNA recombination occurred between the vesicle donor DNA and the chromosomal DNA of the recipient (18). Since no homologous sequences were flanked with drug-resistant genes in the recipient bacterium Modesto (which was confirmed by our genomic sequencing data) in our study, the ARGs could not have been incorporated into the chromosomes of recipient cells through homologous recombination (18). Moreover, even if homologous sequences were present, the likelihood of gene recombination between vesicle donor DNA and recipient chromosomal DNA is low and not as effective as the transfer of genes by plasmids, which has been demonstrated in multiple HGT studies (15, 16). Conclusion. In this study, we present the complete genome sequencing data of A. paragallinarum P4chr1 and its OMVs and confirm that they are highly homologous. In addition, we identified some drug resistance genes in the A. paragallinarum P4chr1 genome that were not present in the A. paragallinarum Modesto genome. Using purified OMVs from P4chr1 as the vector, four AGRs were transferred into the drug-sensitive A. paragallinarum Modesto strain. However, the ARG transformation efficiency and persistency were limited. More studies are required to further understand OMV-mediated HGT with chromosomal DNA-based ARGs.
A. paragallinarum P4chr1 was isolated from the infraorbital sinus sample of a diseased bird from a chicken farm in China in 2021. It was identified as serovar A earlier with a conventional hemagglutination-hemagglutination inhibition (HA-HI) test (20). A. paragallinarum serovar-specific antisera against reference strains 0083 (antiserum A), 0222 (antiserum B), and Modesto (antiserum C) were prepared previously in this laboratory. 16S rRNA gene sequencing and biochemical analyses were used to identify the bacterial species. A. paragallinarum Modesto is a serovar C reference strain preserved in the laboratory (GenBank accession number CP086713).
The isolation and purification steps for OMVs were modified from a previously published procedure (30). Briefly, A. paragallinarum strain P4chr1 was inoculated in tryptic soy broth (TSB) containing 10% (vol/vol) chicken serum and 0.0025% (wt/vol) NAD and cultured for 15 h at 37°C and 180 rpm. The cultured liquid was centrifuged for 30 min at 7,500 × g at 4°C, and the supernatant was filtered through a Stericup filter (Millipore Corporation, Massachusetts, USA) with a pore diameter of 0.45 μm to remove bacteria suspended in the broth. The filtered supernatant was centrifuged for 3 h at 150,000 × g at 4°C (SW40 Ti rotor; Beckman-Coulter, Germany). Then, the supernatant was discarded, and the precipitate was resuspended in 30 mL 0.05 mol/L Tris-HCl buffer (pH 8.0). This process was repeated, and the pellet was resuspended in 5 mL phosphate-buffered saline (PBS) to obtain crude OMVs. The extracted OMVs were centrifuged for 1 h at 75,000 × g at 4°C. Finally, purified OMVs were obtained by resuspending the collected pellet in 2 mL 50 mM HEPES-150 mM NaCl solution.
DNA was extracted as previously described (20) and sequenced via second-generation sequencing methods (Allwegene Technologies, China).
The vesicle suspension was fixed in cold 2.5% (vol/vol) glutaraldehyde for 2 h at 4°C, followed by 1% (wt/vol) osmium tetroxide for 1 h at 4°C. After washing with deionized water, the immobilized vesicles were placed on a 200-mesh grid and imaged using a Philips CM 100 transmission electron microscope (TEM) at 80 kV.
Bacterial genomic DNA was extracted using the Invitrogen DNA minikit (Thermo Fisher Scientific, USA). A. paragallinarum P4chr1 was subjected to WGS using a combination of Nanopore PromethION (Oxford Nanopore Technologies, Beijing, China) and Illumina NovaSeq 6000 (Genewiz, Beijing, China) platforms. Canu v1.5 and Falcon v0.3.0 were used to perform mixed assembly of the original data. The second-generation sequencing-derived small fragment data were used to perform single-base correction (GATK) on the assembly to obtain a high-confidence assembly sequence. Gene prediction was performed using Prodigal software (PROkaryotic DYnamic programming Gene-finding ALgorithm), since Prodigal has high-quality gene structure prediction and better translation initiation site prediction, and gives fewer false positives than other software (31). Prodigal, Glimmer, and GeneMark.hmm were used for gene prediction of NCBI (National Center for Biotechnology Information) prokaryotes. Genomes were annotated using the online database RAST (http://rast.nmpdr.org/), and the results were corrected using the BLASTn database (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The ResFinder database was used to detect ARGs in the genome (https://cge.cbs.dtu.dk/services/). Gene function and metabolic pathway predictions were obtained using the Blast2GO annotation pipeline. The BRIG (BLAST Ring Image Generator) tool was used to draw the circular map of A. paragallinarum P4chr1 and compare it with the OMVs. Transmembrane domains (TMDs) in the P4chr1 genome were predicted using TMHMM Server v.2.0. Finally, the software NCBI BLAST+ was used to compare amino acid sequences of the proteins with the data from the COG, KEGG, VFDB, Nt, Nr, and Swiss-Prot databases to obtain the protein function annotation information.
A. paragallinarum Modesto was grown in 10 mL TSB with the supplements described earlier for 15 h at 37°C and 180 rpm. The 2% (vol/vol) culture was then transferred to 500 mL TSB containing the same supplements and incubated to an optical density (OD) of 0.4 to 0.6. Next, the culture was centrifuged for 30 min at 4,000 × g at 4°C, and the supernatant was discarded. The precipitate was resuspended in 10 mL precooled 272 mM sucrose solution. The culture was centrifuged for 30 min at 4,000 × g, and then the precipitate was resuspended in 1 mL cooled 10% (vol/vol) glycerol and divided into 200-μL aliquots.
The transformation experiment was performed as described previously by Fulsundar et al. (14) (Fig. 8). Experiments were conducted in triplicate three independent times. To prepare the gene transfer incubation mixture, 50 μL of recipient Modesto cells were added to 500 μL super optimal broth with catabolite repression (SOC; 10% [vol/vol] chicken serum and 0.0025% [wt/vol] NAD) medium supplemented with chicken serum and NAD in each Eppendorf tube, followed by addition of 500 μL purified OMVs with known protein concentration (measured with a Bradford assay kit and adjusted to 1 mg/mL). Next, 1 μL 100 μg/μL DNase (final concentration 100 ng/mL) (Thermo Fisher Scientific, California, USA) was added. The tubes were statically incubated for 1 h at 37°C, and then the mixture was transferred aseptically to culture tubes and incubated for a further 2 h with shaking at 150 rpm. Then, 2 mL SOC medium was added, and samples were incubated with shaking for an additional 21 h. Next, the cells were pelleted by centrifugation, resuspended in 1 mL SOC medium, and plated on TSA plates with or without antibiotics. The four test group cells were plated on TSA with four different antibiotics: chloramphenicol, erythromycin, tetracycline, or streptomycin. Three control groups were also prepared. Control A was used to determine the number of receptor cells (CFU/0.1 mL). Control A cells were prepared in the same way as the test groups but did not contain DNase in their sample mixture and were 10 times serial diluted for viable counting before being plated on TSA plates in the absence of antibiotics. Control B was used to assess the effect of DNase on recipient cells. Control B cells were prepared in the same way as the test groups but were plated on antibiotic-free TSA plates. Control C was the negative control. Control C contained only recipient cells and was plated on TSA plates containing the four antibiotics used in the test groups. The plates were incubated for 2 days at 37°C and then evaluated by counting the number of colonies or transformed colonies grown on each plate for every group. The transformation frequency was calculated as the number of transformed colonies over the number of recipient cells.
Based on the P4chr1 genomic information and NCBI ARG sequence (WP_001089068, WP_032491311, WP_000214125, and WP_010890156), four pairs of ARG primers were designed and synthesized (Table 4). The PCR system was 50 μL. The amplification steps were as follows: 95°C for 5 min; 30 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 1 min; and a final step at 72°C for 10 min. PCR products were analyzed by agarose gel electrophoresis. The PCR products were purified and sent to a company for cloning and sequencing (Sangon, Shanghai, China). The sequence information was acquired by aligning the results with sequences obtained from GenBank using BLAST (www.ncbi.nlm.nih.gov/BLAST/). The donor bacterium P4chr1 containing DNA with ARGs was used as the positive control, and water was used as the negative control.
The classical hemagglutination-hemagglutination inhibition (HA-HI) test was conducted as described previously (20). HA antigens were prepared from TSB cultures seeded with acquired transformed colonies, donor strain P4chr1 (serovar A), and recipient strain Modesto (serovar C). PCR-restriction fragment length polymorphism (RFLP) analysis (26) was performed on the recipient bacterium, donor bacterium, and colonies on the resistant plates as described above.
A. paragallinarum P4chr1, Modesto, and transformed colonies were cultured in TSB containing supplements, in the absence of antibiotics. Antimicrobial susceptibility testing was performed using a broth microdilution method according to the protocol described by the Clinical and Laboratory Standards Institute (CLSI) (32). The resultant MIC data were interpreted according to the recommendations outlined in CLSI documents VET08 (32) and M100 (33). E. coli ATCC 29213 served as the quality control strain.
The complete genomic sequence of the chromosomal DNA P4chr1 has been deposited in GenBank under the accession number CP081939. The genomic sequence of its OMVs has been deposited under BioSample number SAMN22170838. We confirm that the data supporting the findings of this study are available within the article, its supplemental materials, and NCBI (GenBank accession number CP081939, BioSample accession number SAMN22170838). | true | true | true |
PMC9603903 | Tahleel Ali-Nasser,Shiri Zayit-Soudry,Eyal Banin,Dror Sharon,Tamar Ben-Yosef | Autosomal dominant retinitis pigmentosa with incomplete penetrance due to an intronic mutation of the PRPF31 gene | 06-10-2022 | Purpose To identify the molecular mechanisms of the development of autosomal dominant retinitis pigmentosa (adRP) with incomplete penetrance in an Israeli Muslim Arab family. Methods Two patients with adRP underwent a detailed ophthalmic evaluation, including funduscopic examination, visual field testing, optical coherence tomography, and electroretinography. Genetic analysis was performed using a combination of whole exome sequencing (WES) and Sanger sequencing. The pathogenicity of the identified intronic variant was evaluated in silico using several web-based tools, in vitro using a minigene-based assay, and in vivo using reverse transcription PCR analysis of lymphocyte-derived RNA. The relative abundance of alternatively spliced transcripts was evaluated using amplicon-based next-generation sequencing. The relative expression levels of PRPF31 and CNOT3 were measured using quantitative PCR (qPCR) analysis. Results The two patients recruited in this study had childhood-onset RP, with night blindness as the initial symptom, followed by concentric restriction of the visual field. The funduscopic findings included narrowed retinal blood vessels and peripheral bone spicule pigmentation. By the third decade of life, the full-field electroretinography findings had been remarkably attenuated. In these patients, we identified a novel heterozygous intronic variant at position +5 of PRPF31 intron 11 (c.1146+5G>T). The same variant was also detected in one asymptomatic family member. Through in silico analysis, the variant was predicted to alter the splicing of intron 11. An in vitro splicing assay and a reverse transcription PCR analysis of lymphocyte-derived RNA revealed that the mutant allele yielded mainly a shorter transcript in which exon 11 was skipped. The skipping of exon 11 was expected to cause a frameshift and an aberrant truncated protein (p.Tyr359Serfs*29). The qPCR analysis revealed reduced PRPF31 expression levels in the mutation carriers, without a significant difference between the affected patient and his asymptomatic brother. We evaluated several factors that have been suggested to correlate with non-penetrance of PRPF31 mutations, including the number of cis-acting MSR1 elements adjacent to the PRPF31 core promoter, CNOT3 expression level, and CNOT3 rs4806718 single-nucleotide polymorphism. None of these factors correlated with non-penetrance in the family in this study. Conclusions We report a novel intronic mutation in PRPF31 underlying adRP. This report expands the spectrum of pathogenic mutations in PRPF31 and further demonstrates the importance of intronic mutations. Moreover, it demonstrates the phenomenon of incomplete penetrance previously associated with PRPF31 mutations. The fact that the non-penetrance in the family in this study could not be explained by any of the known mechanisms suggests the possible contribution of a novel modifier of PRPF31 penetrance. | Autosomal dominant retinitis pigmentosa with incomplete penetrance due to an intronic mutation of the PRPF31 gene
To identify the molecular mechanisms of the development of autosomal dominant retinitis pigmentosa (adRP) with incomplete penetrance in an Israeli Muslim Arab family.
Two patients with adRP underwent a detailed ophthalmic evaluation, including funduscopic examination, visual field testing, optical coherence tomography, and electroretinography. Genetic analysis was performed using a combination of whole exome sequencing (WES) and Sanger sequencing. The pathogenicity of the identified intronic variant was evaluated in silico using several web-based tools, in vitro using a minigene-based assay, and in vivo using reverse transcription PCR analysis of lymphocyte-derived RNA. The relative abundance of alternatively spliced transcripts was evaluated using amplicon-based next-generation sequencing. The relative expression levels of PRPF31 and CNOT3 were measured using quantitative PCR (qPCR) analysis.
The two patients recruited in this study had childhood-onset RP, with night blindness as the initial symptom, followed by concentric restriction of the visual field. The funduscopic findings included narrowed retinal blood vessels and peripheral bone spicule pigmentation. By the third decade of life, the full-field electroretinography findings had been remarkably attenuated. In these patients, we identified a novel heterozygous intronic variant at position +5 of PRPF31 intron 11 (c.1146+5G>T). The same variant was also detected in one asymptomatic family member. Through in silico analysis, the variant was predicted to alter the splicing of intron 11. An in vitro splicing assay and a reverse transcription PCR analysis of lymphocyte-derived RNA revealed that the mutant allele yielded mainly a shorter transcript in which exon 11 was skipped. The skipping of exon 11 was expected to cause a frameshift and an aberrant truncated protein (p.Tyr359Serfs*29). The qPCR analysis revealed reduced PRPF31 expression levels in the mutation carriers, without a significant difference between the affected patient and his asymptomatic brother. We evaluated several factors that have been suggested to correlate with non-penetrance of PRPF31 mutations, including the number of cis-acting MSR1 elements adjacent to the PRPF31 core promoter, CNOT3 expression level, and CNOT3 rs4806718 single-nucleotide polymorphism. None of these factors correlated with non-penetrance in the family in this study.
We report a novel intronic mutation in PRPF31 underlying adRP. This report expands the spectrum of pathogenic mutations in PRPF31 and further demonstrates the importance of intronic mutations. Moreover, it demonstrates the phenomenon of incomplete penetrance previously associated with PRPF31 mutations. The fact that the non-penetrance in the family in this study could not be explained by any of the known mechanisms suggests the possible contribution of a novel modifier of PRPF31 penetrance.
Retinitis pigmentosa (RP), having a worldwide prevalence of approximately 1 in 4,000 individuals, is the most common form of inherited retinal dystrophy (IRD). In RP (also known as rod-cone degeneration), rod photoreceptors are initially more severely affected than cone photoreceptors; thus, the first clinical symptoms are usually night blindness and gradual restriction of the visual field. Ophthalmologic findings include characteristic pigmentation of the midperipheral retina, attenuation of retinal arterioles, and optic disc pallor [1]. RP is one of the most genetically heterogeneous conditions in humans and can be inherited as autosomal dominant (AD), autosomal recessive, or X-linked. To date, more than 90 genes have been implicated in nonsyndromic RP, of which at least 31 are associated with an AD mode of inheritance (RetNet [Retinal Information Network]). Pre-mRNA processing factor 31 (PRPF31) is a ubiquitous pre-mRNA splicing factor that is part of the U4/U6*U5 tri-small nuclear ribonucleoprotein complex of the spliceosome. The heterozygous mutations of PRPF31 are the major causes of adRP (RP11; OMIM No. 600138) [2], accounting for 6%–11% of cases in various populations [3]. Most of more than 130 PRPF31 mutations reported to date are presumed loss-of-function variants, including frameshift, splice site, and nonsense or large-scale insertions or deletions [3]. Genotype-phenotype correlation was observed between the mutation type and age of onset. The lowest age of onset is associated with nonsense and frameshift variants, followed by large deletions or splice variants. In-frame duplications, insertions, or missense variants show the highest median age of onset [3]. The abundance of loss-of-function mutations in patients with RP11, including complete gene deletions, has led to a consensus view that haploinsufficiency is the disease mechanism in this form of RP [4]. An interesting observation (although not statistically significant) is that patients with large-scale deletions (partial or complete gene loss) have a higher age of diagnosis than patients with other mutation types [3]. On the basis of this observation, it could be postulated that an element of dominant negative effect is involved in cases of nonsense, frameshift, indel, in-frame, and missense variants compared with large deletions. This is further supported by recent works that have shown that at least some PRPF31 mutations are associated with a combined haploinsufficiency and dominant-negative disease mechanism [3,5]. A hallmark of PRPF31-associated RP is incomplete penetrance; that is, obligate carriers may be totally asymptomatic but can still pass pathogenic mutations to their offspring. This phenomenon has been reported worldwide in multiple families harboring various PRPF31 mutations [3]. A major determinant of PRPF31 mutation penetrance is the expression level of the nonmutant PRPF31 allele. High expressors of this allele compensate for the deficiency of the mutant allele and have normal vision, whereas low expressors develop RP due to an insufficient amount of functional PRPF31 protein [6]. Wild-type (wt) PRPF31 transcript abundance is a highly variable and heritable characteristic [7]. Several factors contributing to the varied expression levels have been proposed, including expression quantitative trait loci (eQTLs) on chromosome 14, in trans with PRPF31 [7]; the variable expression level of CNOT3, encoding a subunit of the Ccr4-not transcription complex, which binds to and represses the transcription of the PRPF31 promoter [8,9]; and the number of cis-acting minisatellite repeat (MSR1) elements adjacent to the PRPF31 core promoter, which determines the level of PRPF31 transcriptional repression [10,11]. Based on these observations, the mechanism of incomplete penetrance in RP11 has been described as “variant haploinsufficiency,” in which the existence and/or the severity of disease depends on the type of the mutant allele inherited [3], and the levels at which this allele and the wt allele are expressed [4]. Herein, we describe a family segregating adRP with incomplete penetrance due to an intronic mutation of the PRPF31 gene.
In this study, the tenets of the Declaration of Helsinki were followed. This study was approved by the institutional review boards of the participating medical centers, and written informed consent was obtained from all participants. Ophthalmic examination included measurement of the best-corrected visual acuity, visual field testing, slit-lamp biomicroscopy of the anterior segment and ophthalmoscopic examination after pharmacological pupillary dilatation, spectral-domain optical coherence tomography (SD-OCT), and full field electroretinography (ff-ERG).
Genomic DNA was extracted from venous blood samples using a high-salt solution in accordance with the standard protocol [12]. Whole exome sequencing (WES) of Subject II:1 was performed at 3billion (Seoul, South Korea) using xGen Exome Research Panel v2 (Integrated DNA Technologies, Coralville, IA) and Novaseq 6000 (Illumina, San Diego, CA). Sequence reads were aligned to the reference human genome (GRCh37/hg19). Variants were called via the Franklin web-based pipeline by Genoox. The genotyping of family members was performed using PCR amplification of the relevant DNA segments with specific primers (Appendix 1), followed by Sanger sequencing.
An in vitro evaluation was performed using a minigene-based assay. To create wt and mutant minigene constructs, a 1,028-bp DNA fragment harboring PRPF31 exons 10, 11, and 12, and the introns between them was PCR amplified from the genomic DNA of the patient (Appendix 1). Fragments harboring either the wt or mutant allele were inserted in the pCMV-Script mammalian expression vector (Stratagene, La Jolla, CA). Constructs were transfected into COS-7 cells using the jetPEI transfection reagent (Polyplus-transfection, Illkrich, France). Cells were cultured in a DMEM culture medium supplemented with 10% fetal bovine serum (Biologic Industries, Beit Ha'emek, Israel) and maintained at 37 °C and 5% CO2. Twenty-four hours after transfection, total RNA was extracted from the cells using TRI Reagent (Sigma-Aldrich, St Louis, MO) and treated with RQ1 RNase-free DNase (Promega, Madison, WI). Reverse transcription was performed with 1 μg of DNase-treated total RNA in a 20-μl reaction volume using 200 U of M-MLV reverse transcriptase and 100 ng of random primers (Stratagene). Two microliters of cDNA were subjected to PCR amplification with a forward primer in exon 10 and a reverse primer in the vector (T7; Appendix 1). The PCR products were subcloned into the pGEM-Teasy vector (Promega), and several independent clones of each product were sequenced. The COS-7 cells used for this experiment were authenticated by genotyping 4 STR markers (D17S1304, D5S1467, D4S2408, and D19S245). Genotyping was performed with PCR amplification of each STR and direct sequencing, as described by Almeida et al. [13]. The results (in terms of the repeat number for each STR) were compatible with those described for COS-7 cells (data not shown) [13].
For the in vivo evaluation, total RNA was isolated from fresh blood lymphocytes using TRI Reagent, treated with RQ1 RNase-free DNase, and reverse transcribed as described earlier. PCR amplification was performed with forward and reverse primers in exons 10 and 12, respectively (Appendix 1). For sequencing, the PCR products were cloned into the pGEM-Teasy vector (Promega).
Amplicon-based next-generation sequencing (NGS) was performed at the Technion Genomics Center. Lymphocyte-derived RT-PCR products (described in the previous section, “In vivo splicing analysis”) were prepared for NGS according to the “16S Library Preparation protocol” by Illumina. This included PCR product purification, barcode attachment, and second purification of the final libraries. The concentration of each library was measured using the Qubit dsDNA HS Assay Kit (Life Technologies, Thermo Fisher Scientific, Waltham, MA), and the size was determined using TapeStation 4200 with the High Sensitivity D1000 Kit (Agilent, Santa Clara, CA). All libraries were mixed into a single tube with equal molarity. DNaseq data were generated on Illumina MiSeq as 150 paired-end reads. Demultiplexing was executed using bcl2fastq version v2.20.0.422 software (Illumina), with one barcode mismatch allowed and a minimum trimmed read length of 35. Quality control was assessed using Fastqc (v0.11.8), 150-bp paired-end reads were trimmed for adapters, low-quality 3′, and a minimum length of 30 using CUTADAPT (v1.10). Next, the reads were simultaneously mapped to two references: long.fasta, including exons 10, 11, and 12, and short.fasta, including exons 10 and 12. Mapping was performed using the BBSplit tool from BBTools (version 38.95) with default parameters, where ambiguous reads were binned to the first best reference.
cDNA was synthesized using the qPCRBIO High-Quality cDNA Synthesis Kit (PCR Biosystems, London, UK). qPCR reactions were prepared using 2X qPCRBIO Fast qPCR SyGreen Blue Mix (PCR Biosystems, London, UK) and performed using the QuantStudio3 Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific). Data were analyzed using the ∆∆CT method. Gene expression values were normalized against the GAPDH expression level. The primer sequences are presented in Appendix 1.
TB209 is a Muslim Arab family from Northern Israel, segregating adRP. The affected individuals (subjects II:1 and III:1; Figure 1A) were a mother and her son, aged 54 and 32 years old at the time of the last follow-up, respectively. Both reported childhood-onset night blindness as the initial symptom, which was associated with subjectively worsening dark and light adaptation, followed by a progressive, concentric restriction of the visual field. Subject II:1 also complained of decreasing visual acuity in both eyes several years prior. A posterior subcapsular cataract was present in both eyes of both subjects. The ophthalmoscopic findings included pallor of the optic nerve head, cellophane macular reflex, and narrowed retinal arteries with perivascular and midperipheral bone spicule pigmentation of the retina. Macular imaging using SD-OCT revealed loss of the outer nuclear layer (ONL) and inner segment ellipsoid zone (EZ) band, with preservation of this landmark only in the foveal center in Subject II:1. This patient also had focal vitreomacular traction with cystoid macular changes mainly in the right eye. In Subject III:1, the SD-OCT imaging revealed relative preservation of the EZ and ONL in the fovea, whereas ff-ERG at the age of 29 years was attenuated to a non-measurable level (Table 1 and Figure 2).
To identify the genetic cause of the disease in family TB209, WES analysis (including copy number variation detection) was performed on a DNA sample from Subject II:1. A heterozygous variant at position +5 of intron 11 in the PRPF31 gene (NM_ 015629) was identified (c.1146+5G>T; IVS11+5G>T; Figure 1B, C). Through in silico analysis, this variant was predicted to alter the splicing of intron 11 (Table 2). The same variant appeared to be heterozygous in Subject III:1 (affected) and in his unaffected brother (Subject III:3; Figure 1A). This novel variant has never been reported in patients with IRD (neither in the literature nor in ClinVar); is not registered in public databases, including the Genome Aggregation Database, Trans-Omics for Precision Medicine (TOPMed) Bravo, The Greater Middle East (GME) Variome, and the Exome Sequencing Project (ESP) 6500; and is absent in more than 1000 exomes of Israeli patients with IRD.
We initially used an in vitro splicing assay approach that enabled us to examine each allele (wt vs. mutant) separately and evaluate the effect of c.1146+5G>T on splicing. For this purpose, we created two minigene constructs (wt and mutant) harboring PRPF31 exons 10 to 12 and the introns between them, downstream of a CMV promoter. Constructs were transfected into COS-7 cells, followed by RNA extraction and RT-PCR analysis. To specifically detect PRPF31 transcripts obtained from the minigene constructs (and not from the endogenously expressed gene), we performed PCR amplification with a forward primer derived from exon 10 and a reverse primer derived from the expression vector (Figure 3A). The wt construct yielded both the expected product harboring exons 10, 11, and 12 (exon 11+) and a shorter product in which exon 11 was skipped (exon 11−). The mutant construct mainly yielded the shorter product (exon 11−; Figure 3A,B). The skipping of exon 11 was expected to cause a frameshift that yields an aberrant truncated protein (p.Tyr359Serfs*29). To evaluate the effect of c.1146+5G>T on splicing in vivo, we performed an RT-PCR analysis of leukocyte-derived RNA samples for all available family members, with primers located in exons 10 and 12. As expected, exon 11+ transcripts were detected in all tested individuals. Exon 11− transcripts were mainly detected in the individuals heterozygous for the c.1146+5G>T allele, although faint products could also be observed in the wt individuals (Figure 4A). To quantify the amount of exon 11+ relative to that of exon 11− transcripts in the wt and heterozygous individuals, we applied amplicon-based NGS to the RT-PCR products of family members. The quantitative analysis of the NGS results confirmed that skipping of exon 11 occurred in both the wt and mutant individuals. Exon 11− reads composed 2%–5% of all unique reads in the wt individuals and 8%–10% in the individuals heterozygous for the c.1146+5G>T allele (Figure 4B, C).
As mentioned earlier, Subject III:3 was heterozygous for the c.1146+5G>T mutation. The clinical data of this individual were not available, but at the age of 32 years, he subjectively reported good vision and indicated that he was examined by an ophthalmologist, with no significant findings. Considering that both his mother and brother experienced onset of night blindness in childhood and that his brother's ff-ERG was non-recordable at the age of 29 years, it is unlikely that he was affected, although late onset of disease symptoms cannot be excluded. One factor that affects the penetrance of PRPF31 mutations is the expression level of the nonmutant PRPF31 allele [6]. CNOT3 negatively regulates the expression of PRPF31, and in some studies, the lymphoblasts of asymptomatic PRPF31 mutation carriers were found to have higher PRPF31 expression levels and lower CNOT3 expression levels than those of symptomatic carriers [8,9]. Hence, we evaluated several factors that have been suggested to correlate with the non-penetrance of PRPF31 mutations. One was the CNOT3 rs4806718 single-nucleotide polymorphism (SNP; C/T). A correlation between the T allele of this polymorphism and non-penetrance was previously suggested [9]. In family TB209, the CNOT3 rs4806718 genotype did not correlate with disease penetrance, as all three carriers of the PRPF31 mutation, both symptomatic (subjects II:1 and III:1) and asymptomatic individuals (Subject III:3), were heterozygotes for this SNP. We also determined the number of cis-acting MSR1 elements adjacent to the PRPF31 core promoter; four-copy repeats have been associated with non-penetrance [10,11]. In family TB209, no such correlation was found, as all family members, including the non-penetrant individual, were homozygous for the three-copy MSR1 repeat. We then evaluated the relative expression levels of PRPF31 and CNOT3 in lymphocyte samples from TB209 family members using qPCR. The PCR primers used for PRPF31 amplification were located in exons 3 and 5, and therefore amplified all PRPF31 transcripts, including both the wt and mutant. However, given the low abundance of exon 11− transcripts (Figure 4), most amplification products were expected to represent wt transcripts. The analysis revealed that subjects III:1 (symptomatic) and III:3 (asymptomatic) had similar and reduced expression levels of PRPF31 (approximately 50% of the average expression level in the wt controls; Figure 5). The CNOT3 expression level showed no statistically significant difference between the two individuals and was similar to the average expression level in the wt controls (Figure 5). Finally, we genotyped the family members for three SNPs in the PRPF31 gene to reconstruct the PRPF31-linked haplotype in chromosomes harboring the wt rather than the mutant allele. The analysis revealed that subjects III:1 and III:3 inherited different copies of chromosome 19 from their father (Appendix 2). These results support the presence of a genetic modifier leading to the non-penetrance in Subject III:3 in cis with PRPF31, although the effect of such a putative modifier on the PRPF31 expression level was not apparent in the lymphocytes.
The primary aim of the present study was to identify the molecular mechanism of adRP in an Israeli Muslim Arab family. Genetic analysis revealed a novel heterozygous intronic mutation of PRPF31, c.1146+5G>T. The affected individuals carrying this mutation had classic symptoms of RP with childhood onset. The clinical findings from these patients were similar to those reported previously in patients with other PRPF31 mutations [3]. To evaluate the effect of c.1146+5G>T on splicing, we performed in vitro and in vivo splicing analyses. The results of the analyses indicated that some degree of alternative splicing of exon 11 is normal and that the c.1146+5G>T allele further weakens the intron 11 donor splice site and enhances the skipping of intron 11. At the protein level, skipping of exon 11 was expected to cause a frameshift that yields an aberrant truncated protein (p.Tyr359Serfs*29). The presence of such a truncated protein was not confirmed experimentally. If it was indeed generated, it could be associated with either loss of function and/or a dominant negative effect [3]. Nevertheless, although exon 11− transcripts were more abundant in the individuals harboring the c.1146+5G>T allele than in the wt individuals, the relative frequency of these transcripts was low in both groups. Moreover, the total PRPF31 expression level in the carriers of the c.1146+5G>T allele was approximately half of the expression level in the wt individuals. These results indicate that the mutant transcript was probably unstable and may be subjected to the NMD mechanism and support haploinsufficiency as the disease mechanism associated with the c.1146+5G>T mutation. In this context, it is of special interest to mention another PRPF31 mutation, c.1115_1125del. This mutation is located in exon 11 and can inactivate an exonic splicing enhancer and thus promote the skipping of exon 11 during mRNA splicing and result in an out-of-frame premature termination codon in exon 12, similar to the effect of c.1146+5G>T [14]. Exon 11− transcripts generated as a result of this mutation were shown to undergo degradation through the NMD mechanism [14]. One family member carried the c.1146+5G>T mutation with no phenotypic expression. This phenomenon has been previously reported in families segregating PRPF31 mutations [3]. However, we could not identify the factors contributing to the reduced penetrance in this family. The CNOT3 rs4806718 genotype did not correlate with disease penetrance in this family, which is in agreement with a previous study [10] that did not replicate the original observation made by Venturini et al. [9]. RP11 disease non-penetrance was also associated with the presence of a four-copy MSR1 repeat in the PRPF31 promoter in many RP11 asymptomatic carriers [10,11]. In family TB209, all family members, including the asymptomatic carrier, were homozygous for the three-copy MSR1 repeat. PRPF31 penetrance correlates, at least in part, with the expression level of the nonmutant PRPF31 allele [6], which reversely correlated with the CNOT3 expression level [8,9]. In family TB209, the asymptomatic carrier had a reduced PRPF31 expression level, which was similar to the expression level in his affected brother. The CNOT3 expression levels were similar in both individuals and the wt controls. However, these expression levels were measured in blood lymphocytes and may not reflect the expression levels in retinal photoreceptors. For example, in a study performed by McLenachan et al. [10], the PRPF31 expression level of a non-penetrant carrier was not higher than that of symptomatic individuals when measured in skin fibroblasts but was an intermediate between wt and symptomatic individuals when measured in retinal organoid cultures (differentiated from induced pluripotent stem cells). It is possible that the asymptomatic carrier in family TB209 had an increased expression level of the wt PRPF31 allele in his retina, which compensated for the mutant allele. Alternatively, the non-penetrance in this individual might have resulted from other genetic and/or environmental factors that are still to be determined. In conclusion, this report expands the spectrum of pathogenic mutations in PRPF31 and further demonstrates the importance of intronic mutations. Moreover, it demonstrates the phenomenon of incomplete penetrance, which was previously associated with PRPF31 mutations. The fact that non-penetrance in the family in this study could not be explained by any of the known mechanisms suggests the possible contribution of a novel modifier of PRPF31 penetrance. | true | true | true |
||
PMC9603908 | 36173294 | Yuhang Zhang,Qingmei Li,Ruining Wang,Li Wang,Xun Wang,Jun Luo,Guangxu Xing,Guanmin Zheng,Bo Wan,Junqing Guo,Gaiping Zhang | Differentiation of Classical Swine Fever Virus Virulent and Vaccine Strains by CRISPR/Cas13a | 29-09-2022 | classical swine fever virus,recombinase-aided amplification,CRISPR/Cas13a,differentiation of infected and vaccinated animals | ABSTRACT As a notifiable terrestrial and aquatic animal disease listed by World Organisation for Animal Health (formerly the Office International des Epizooties [OIE]), classical swine fever (CSF) has caused great economic losses to the swine industry worldwide during recent decades. Differentiation of infected and vaccinated animals (DIVA) is urgent for eradication of CSF. In this study, a diagnostic platform based on CRISPR/Cas13a was established with the ability to differentiate between classical swine fever virus (CSFV) virulent and vaccine strains. In combination with reverse transcription recombinase-aided amplification (RT-RAA), the detection limit for CSFV synthetic RNA templates reached 3.0 × 102 copies/μL. In addition, with boiling and chemical reduction, heating unextracted diagnostic samples to obliterate nucleases (HUDSON) treatment was introduced to inactivate nucleases and release viral genome, achieving robust pretreatment of tested sample before CRISPR/Cas13a detection without the need to extract viral nucleic acids. HUDSON-RT-RAA-CRISPR/Cas13a can directly detect cell cultures of virulent Shimen strain and vaccine hog cholera lapinized virus (HCLV) strain, with the detection limit of 3.5 × 102 copies/μL and 1.8 × 102 copies/μL, respectively, which was equally sensitive to nested PCR (nPCR) and 100 times more sensitive than antigen enzyme-linked immunosorbent assay (ELISA). Meanwhile, HUDSON-RT-RAA-CRISPR/Cas13a showed no cross-reactivity with bovine viral diarrhea virus (BVDV), atypical porcine pestivirus (APPV), porcine reproductive and respiratory syndrome virus (PRRSV), porcine epidemic diarrhea virus (PEDV), African swine fever virus (ASFV), pseudorabies virus (PRV), and porcine circovirus 2 (PCV2), exhibiting good specificity. At last, a total of 50 pig spleen samples with suspected clinical signs were also assayed with HUDSON-RT-RAA-CRISPR/Cas13a, nPCR, and antigen ELISA in parallel. HUDSON-RT-RAA-CRISPR/Cas13a showed 100.0% with nPCR and 82.0% coincident rate with antigen ELISA, respectively. IMPORTANCE Classical swine fever (CSF) is a World Organisation for Animal Health (formerly the Office International des Epizooties [OIE]) notifiable terrestrial and aquatic animal disease, causing great economic losses to the swine industry worldwide during the past decades. Due to the use of the most effective and safe attenuated live vaccine for CSF prevention, differentiation of infected and vaccinated pigs is vital work, as well as a bottleneck for eradication of CSF. Methods with the ability to precisely differentiate classical swine fever virus (CSFV) virulent strains from vaccine strain hog cholera lapinized virus (HCLV) are urgently needed. Combining the high sensitivity of isothermal recombinase-aided amplification (RAA) with the accurate molecular sensing ability of Cas13a, we presented a novel method for CSFV detection without the need to extract viral nucleic acids, which showed great advantage to traditional detection methods for precise differentiation of CSFV virulent strains and vaccine strain, providing a novel powerful tool for CSF eradication. | Differentiation of Classical Swine Fever Virus Virulent and Vaccine Strains by CRISPR/Cas13a
As a notifiable terrestrial and aquatic animal disease listed by World Organisation for Animal Health (formerly the Office International des Epizooties [OIE]), classical swine fever (CSF) has caused great economic losses to the swine industry worldwide during recent decades. Differentiation of infected and vaccinated animals (DIVA) is urgent for eradication of CSF. In this study, a diagnostic platform based on CRISPR/Cas13a was established with the ability to differentiate between classical swine fever virus (CSFV) virulent and vaccine strains. In combination with reverse transcription recombinase-aided amplification (RT-RAA), the detection limit for CSFV synthetic RNA templates reached 3.0 × 102 copies/μL. In addition, with boiling and chemical reduction, heating unextracted diagnostic samples to obliterate nucleases (HUDSON) treatment was introduced to inactivate nucleases and release viral genome, achieving robust pretreatment of tested sample before CRISPR/Cas13a detection without the need to extract viral nucleic acids. HUDSON-RT-RAA-CRISPR/Cas13a can directly detect cell cultures of virulent Shimen strain and vaccine hog cholera lapinized virus (HCLV) strain, with the detection limit of 3.5 × 102 copies/μL and 1.8 × 102 copies/μL, respectively, which was equally sensitive to nested PCR (nPCR) and 100 times more sensitive than antigen enzyme-linked immunosorbent assay (ELISA). Meanwhile, HUDSON-RT-RAA-CRISPR/Cas13a showed no cross-reactivity with bovine viral diarrhea virus (BVDV), atypical porcine pestivirus (APPV), porcine reproductive and respiratory syndrome virus (PRRSV), porcine epidemic diarrhea virus (PEDV), African swine fever virus (ASFV), pseudorabies virus (PRV), and porcine circovirus 2 (PCV2), exhibiting good specificity. At last, a total of 50 pig spleen samples with suspected clinical signs were also assayed with HUDSON-RT-RAA-CRISPR/Cas13a, nPCR, and antigen ELISA in parallel. HUDSON-RT-RAA-CRISPR/Cas13a showed 100.0% with nPCR and 82.0% coincident rate with antigen ELISA, respectively. IMPORTANCE Classical swine fever (CSF) is a World Organisation for Animal Health (formerly the Office International des Epizooties [OIE]) notifiable terrestrial and aquatic animal disease, causing great economic losses to the swine industry worldwide during the past decades. Due to the use of the most effective and safe attenuated live vaccine for CSF prevention, differentiation of infected and vaccinated pigs is vital work, as well as a bottleneck for eradication of CSF. Methods with the ability to precisely differentiate classical swine fever virus (CSFV) virulent strains from vaccine strain hog cholera lapinized virus (HCLV) are urgently needed. Combining the high sensitivity of isothermal recombinase-aided amplification (RAA) with the accurate molecular sensing ability of Cas13a, we presented a novel method for CSFV detection without the need to extract viral nucleic acids, which showed great advantage to traditional detection methods for precise differentiation of CSFV virulent strains and vaccine strain, providing a novel powerful tool for CSF eradication.
Classical swine fever (CSF) is a highly contagious and often fatal infectious porcine disease, leading to high fever, multisystemic hemorrhagic lesions, and immunosuppression in infected pigs. It has brought great economic losses to the swine industry worldwide during recent decades and is thus classified as one of notifiable terrestrial and aquatic animal diseases by the World Organisation for Animal Health (formerly the Office International des Epizooties [OIE]) (1). Classical swine fever virus (CSFV), the causative agent of CSF, belongs to the family Flaviviridae, genus Pestivirus. It is a small enveloped virus with a single-stranded, positive-sense RNA genome. The genomic RNA contains a large open reading frame flanked with a 5′ untranslated region (5′UTR) and a 3′ untranslated region (3′UTR), encoding four structural proteins (C, Erns, E1, and E2) and eight nonstructural proteins (Npro, p7, NS2, NS3, NS4A, NS4B, NS5A, and NS5B) (2). Based on partial sequences of 5′UTR, E2 and NS5B, CSFV can be divided into three genotypes and 11 subgenotypes (1.1 to 1.4, 2.1 to 2.3, and 3.1 to 3.4) (3). Since the effective and safe attenuated live vaccine hog cholera lapinized virus (HCLV) strain was produced by serial passage of CSFV Shimen strain in rabbits at 1956, many countries, including China, use a systematic prophylactic vaccination policy to control or eradicate CSF, excepting for some European and American countries that carry out a stamping out policy (nonvaccination policy). To date, pandemics of CSF have been effectively prevented with either vaccination or nonvaccination policies. Instead, sporadic and occasional epidemics of large-scale outbreaks have become current epidemiological trend of CSF (4, 5). Since 2017, Chinese government has begun to take a series of actions for CSF eradication, including vaccination and diagnostic policies. One obstacle of CSF elimination is that both the virulent Shimen strain and vaccine HCLV strain belong to the same serotype and subgenotpye 1.1, making it hard for traditional immunological or molecular diagnostic methods to differentiate infected pigs from vaccinated ones (6–10). Although efforts have been made to develop alternative subunit or gene-deleted vaccines for differentiation of infected and vaccinated animals (DIVA) purposes (11), HCLV-based attenuated live vaccine still remains the primary choice for CSF prevention in China. Methods with the ability to precisely differentiate CSFV virulent strains from vaccine strain HCLV are urgently needed. Current studies about CSFV DIVA mainly rely on molecular tests, including sequencing, PCR, and restrictive-fragment-length polymorphism (12, 13). However, these methods are often laborious and time consuming and need to be performed by skilled technicians with expensive instruments, using multiple testing steps and complex reagents. Cas13a is a single-component RNA-guided RNA-targeting CRISPR effector, belonging to type VI-A within CRISPR/Cas family. Guided by CRISPR RNA (crRNA), Cas13a cleaves both the target RNAs and nonspecific single-stranded RNA (ssRNA) (14, 15). The nonspecific ssRNA cleavage activity of Cas13a is called collateral RNase activity and can be assayed with a fluorophore-quencher-labeled RNA substrate (reporter RNA) for fluorescent readout (16). In combination with recombinase-based isothermal amplification and collateral cleavage assay for Cas13a, a promising next-generation diagnostic strategy for infectious diseases was developed and named SHERLOCK for Specific High Sensitivity Enzymatic Reporter UnLOCKing. SHERLOCK is an isothermal method that can detect both DNA and RNA target at an attomolar level with single-base mismatch specificity (17). In addition, Heating Unextracted Diagnostic Samples to Obliterate Nucleases (HUDSON) treatment can lyse viral particles and inactivate ribonucleases with the use of heat and chemical reduction, avoiding extraction of viral nucleic acids. With high performance and minimal sample processing requirements, HUDSON-SHERLOCK is a rapid and convenient diagnostic method that has been used for differentiation of four related flaviviruses viruses, including Zika virus, dengue virus, West Nile virus, and yellow fever virus (18). Combined the high sensitivity of recombinase-aided amplification (RAA) with the accurate molecular sensing ability of Cas13a, in this study, we established a novel DIVA platform to differentiate between CSFV virulent strains and vaccine strain without the need to extract viral nucleic acids (Fig. 1), providing a promising strategy for CSF diagnosis and eradication.
A single band with a molecular weight of about 140 kDa was obtained by SDS-PAGE after nickel affinity chromatography and sulfopropyl (SP) cation exchange chromatography, illustrating that LwaCas13a was successfully expressed in soluble state and purified to a desired degree (Fig. S1). A total of 23.04 mg purified LwaCas13a was obtained from 3-liter bacterial LB cultures.
Among five candidate crRNAs for HCLV strain, crHCLV5 produced the highest fluorescent signals, exhibiting the ability to recognize HCLV 3′UTR RNA and activate LwaCas13a efficiently (Fig. S2A). Similarly, among nine candidate crRNAs for the Shimen strain, crShimen7 produced the highest fluorescent signal, exhibiting the ability to recognize Shimen 3′UTR RNA and activate LwaCas13a efficiently (Fig. S2B).
To estimate the ability of crHCLV5 and crShimen7 to differentiate between CSFV vaccine strain and virulent strains, LwaCas13a collateral cleavage assay was performed to test different genotypes of CSFV 3′UTR RNAs, including subgenotypes 1.1, 1.2, 1,4, 2.2, 2.3, 3.2, and 3.4, as well as subsubgenotypes 2.1a, 2.1b, 2.1c, 2.1d, 2.1g, 2.1h, and 2.1i. For crHCLV5, only the cognate HCLV strain can be recognized, indicating that the vaccine strain can be differentiated from all tested virulent strains by CRISPR/Cas13a-crHCLV5 (Fig. 2A). For crShimen7, only four virulent subgenotypes can be recognized: 1.1, 1.2, 2.2, and 3.2 (Fig. 2B). Considering the dominant pandemic subgenotype in China was 2.1, protospacer sequences of all virulent genotypes were further analyzed. Alignment results showed that there were three main mutations in subgenotype 2.1, including G to A at position 5, U to C at position 16, and U to A at position 19. Based on these mutations, cr2.17 was designed to detect CSFV all subsubgenotypes of 2.1. Collateral cleavage assay showed that even though the efficiencies of cr2.17 to recognize different subsubgenotypes of 2.1 were varied, which depends on the number and position of mutations, cr2.17 can differentiate all subsubgenotypes of 2.1 (including 2.1a, 2.1b, 2.1c, 2.1d, 2.1g, 2.1i, and 2.1h) from the vaccine strain (Fig. 2C). For the purpose of detecting both the traditional virulent Shimen strain and the dominant pandemic virulent strains of subgenotype 2.1, crShimen7 and cr2.17 were both introduced in collateral cleavage assay. The results showed that except for subgenotypes 2.4 and 3.4, CRISPR/Cas13a-crShimen/cr2.17 recognized most virulent subgenotypes, including the traditional Shimen strain and all dominant pandemic virulent subsubgenotypes of 2.1 that have been reported to exist in China. In addition, CRISPR/Cas13a-crShimen7/cr2.17 showed no cross-reactivity to the HCLV strain, exhibiting the ability to differentiate virulent strains from the vaccine strain (Fig. 2D).
A recent study compared two recombinase-based techniques, recombinase polymerase amplification (RPA) and RAA, in parallel for rapid detection for African swine fever virus (ASFV), demonstrating that the detection limit for RPA was 93.4 copies/reaction, while the detection limit for RPA was 53.6 copies/reaction (19). Thus, RAA was introduced to improve detecting sensitivity for this study. The detection limit of CRISPR/Cas13a was estimated with synthetic Shimen and HCLV 3′UTR RNA by collateral cleavage assay. The results showed that CRISPR/Cas13a alone did not yield a detectable signal at input concentrations below 3.0 × 1010 copies/μL (Fig. 3A and B). To enhance detecting sensitivity of CRISPR/Cas13a, RT-RAA was introduced to preamplify tested samples before collateral cleavage assay (Fig. S5; Table S5). The results showed that the sensitivity of RT-RAA-CRISPR/Cas13a was greatly improved compared to CRISPR/Cas13a alone, with a detection limit of 3.0 × 102 copies/μL (Fig. 3C and D).
Before estimating the specificity of RT-RAA-CRISPR/Cas13a, cDNAs common porcine viruses or members of pestiviruses were first tested by nPCR, except for ASFV, which used certified reference genomic nucleic acid material as the template and was tested by OIE-recommended PCR. The results showed that all of the tested viruses can be efficiently amplified (Fig. S4A), confirming the reliability of all samples used in specificity estimation. Collateral cleavage assay showed that when extracted genomic nucleic acids of viral cell cultures were used as the templates, RT-RAA-CRISPR/Cas13a was also able to differentiate between Shimen and HCLV. In addition, RT-RAA-CRISPR/Cas13a showed no cross-reactivity to other viruses, exhibiting good detecting specificity (Fig. 4A and B).
In order to simplify the procedure of RT-RAA-CRISPR/Cas13a detection, the HUDSON technique was introduced for robust treatment of tested samples without the need to extract nucleic acids. To estimate the compatibility of HUDSON and RT-RAA-CRISPR/Cas13a, cell cultures of Shimen and HCLV strains were treated with HUDSON and directly used for collateral cleavage assay. Only slightly decreases of fluorescent signals were observed when HUDSON-treated samples were tested compared with cognate genomic nucleic acids, maintaining a comparable detecting sensitivity (Fig. 5A and B).
Shimen and HCLV cell cultures and 10-fold serial dilutions from 10−1 to 10−4 were tested by HUDSON-RT-RAA-CRISPR/Cas13a, nPCR and antigen enzyme-linked immunosorbent assay (ELISA) in parallel. For both HUDSON-RT-RAA-CRISPR/Cas13a and nPCR, positive results were obtained at a maximum dilution level of 10−1 for HCLV and 10−2 for Shimen (Fig. 6A and B), showing an equal sensitivity to nPCR (Fig. 6C). The detection limits for cell cultures of Shimen and HCLV were 3.5 × 102 and 1.8 × 102 copies/μL, respectively, which were coincident with the detection limits for testing synthetic RNA templates. For antigen ELISA, a positive result was obtained only when testing Shimen cell culture without dilution, while suspected positive results were obtained when testing Shimen cell culture with 10−1 dilution and HCLV cell culture without dilution, indicating that antigen ELISA was at least 100 times less sensitive than HUDSON-RT-CRISPR/Cas13a and nPCR (Fig. 6D). Before comparison of these three methods for testing tissue samples, cDNA of 50 spleen samples was first tested by PCR with individual specific primers of bovine viral diarrhea virus (BVDV), atypical porcine pestivirus (APPV), porcine reproductive and respiratory syndrome virus (PRRSV), porcine epidemic diarrhea virus (PEDV), African swine fever virus (ASFV), pseudorabies virus (PRV), and porcine circovirus 2 (PCV2) to investigate the background of each sample (Fig. S4). The results showed that the infection situations were complex among these samples, some of which exhibited coinfection with more than one virus; 15 samples showed CSFV positive, including samples 22, 35 to 38, and 41 to 50. Unfortunately, nPCR for BVDV used in this study cannot differentiate BVDV from CSFV. Considering that primers for CSFV can only amplify CSFV while primers for BVDV can amplify both BVDV and CSFV, no BVDV positivity was found among the tested samples. Similarly, no APPV-infected sample was found. One sample (sample 27) was found to be PEDV positive, and 23 samples were PRRSV positive, including samples 02 to 06, 13, 16 to 20, 24 to 26, 28 to 34, 39, and 40. Because all of these samples were collected between 2016 and 2018, a period before the first ASF outbreak in China in 2019, no ASFV-positive sample was found. Three samples were PRV positive, including samples 07, 16, and 23. Twenty-seven samples were PCV2 positive, including 01, 03 to 05, 08 to 16, 18 to 21, 23 to 25, 31 to 34, 39, and 40. A total of 15 samples tested positive, and 35 samples were negative by HUDSON-RT-RAA-CRISPR/Cas13a, which showed 100% coincident rate with nPCR. However, due to the lower detecting sensitivity of antigen ELISA, among these 15 CSFV-positive samples, only 6 samples tested positive, while 4 samples were suspected positive, and another 5 samples tested negative by antigen ELISA (Table S6; Fig. S4), showing 82.0% coincident rate (Table 1).
Accurate diagnosis of important infectious diseases has been paid increasing attention in recent years, especially since the outbreaks of ASF and COVID-19 that have brought about enormous economic losses worldwide (20, 21). With the development of science and technology, continuously emerging techniques have been developed to meet specific diagnostic demands of various infectious diseases that traditional methods could not achieve. Due to the impressive ability to precisely recognize nucleic acids, CRISPR/Cas9 has won the Nobel prize as a powerful gene-editing tool for rewriting the code of life (22). Similarly, as another member of the Cas family, Cas13a is also able to precisely recognize RNA with single-base specificity. With the unique collateral RNase activity, Cas13a is first effector that transforms CRISPR technology from gene scissor into gene sensor (17). In this study, we established CRISPR/Cas13a platform for CSFV DIVA application, which is one of the greatest obstacles for CSF eradication. Even though CRISPR/Cas13a alone is not sensitive enough, RT-RAA significantly increased the detecting sensitivity, with a level equal to that of nPCR. Combining the high specificity of CRISPR with the high sensitivity of RAA is a promising platform for differential diagnosis of infectious diseases. Due to the simplicity of primer requirement and base mismatch tolerance of RAA (23), it is easy to design universal primers to detect a broad spectrum of all genotypes of a specific virus like CSFV, followed by differential detection by CRISPR/Cas13a. In addition, both RAA and CRISPR collateral cleavage assay were applied at a constant 37°C, showing the potential to be developed as point-of-care testing tools. By simply heating samples with EDTA and Tris(2-carboxyethyl)phosphine hydrochloride (TCEP), HUDSON treatment can robustly release nucleic acids and inactivate ribonucleases for downstream molecular detection. It can be further combined with commercial lateral flow strips and biotin-6-carboxyfluorescein (FAM)-labeled polyU reporter to replace fluorescent readout with visual readout, achieving complete instrument-free detection from sample processing to diagnostic results (24). In addition to Cas13a, Cas12a is another CRISPR effector with collateral DNase activity that has been widely developed as a gene-sensing tool in recent years (25). Cas12a is able to directly target DNA molecules, it shows advantages beyond Cas13a as a more stable, simpler, and less expensive tool. However, Cas13a has fewer sequential restrictions than Cas12a, with a protospacer flanking site requirement (not G) rather than protospacer adjacent motif requirement (TTTV). In this study, a 50-nucleotide (nt)-long region from the CSFV 3′UTR area contains nine candidate Cas13a crRNAs for Shimen strain and five crRNAs for HCLV strain. While no suitable Cas12a crRNA was found in the same region, Cas13a exhibited little dependency to target sequence than Cas12a. Different characters of these two effectors make Cas13a more suitable for differential diagnosis with the two-step procedure established in this study, while Cas12a is more suitable for one-pot assay (26–28) or amplification-free detection (29–32) that has been presented in other works. With an increasing number of Cas effectors with various characters to be explored and developed as gene-sensing tools, studies about the CRISPR method of diagnosing infectious diseases is a promising direction for the future.
In this study, we established CRISPR/Cas13a platform for differentiation between CSFV virulent and vaccine strains with high sensitivity and good specificity. We also introduced HUDSON treatment to replace nucleic acids extraction step, facilitating the detection procedure and reducing detection time. The established platform can be performed at a constant 37°C, showing the potential to be developed as point-of-care testing. Future work will focus on development of convenient readout methods, such as lateral flow assay or portable fluorescent devices.
Leptotrichia wadei Cas13a (LwaCas13a) bacterial expression plasmid Twinstrep-SUMO-LwaCas13a was purchased from Addgene (https://www.addgene.org/90097/) and transformed into Rosetta (DE3) competent cells. Positive clones were verified by sequencing and preserved as bacterial glycerol stock at −80°C. Complete CSFV 3′UTR of different genotypes (Table S1) with an upstream T7 promoter sequence (5′-TAATACGACTCACTATAGGGG-3′) were synthesized by Sangon Biotech and cloned into PUC57 plasmid. Other DNA oligonucleotides, including primers and crRNA transcription templates, were also synthesized by Sangon Biotech (China). All viral cell cultures and tissue samples were provided by the Key Laboratory of Animal Immunology of Henan Academy of Agricultural Sciences, including the cell culture of CSFV, viral diarrhea virus (BVDV), porcine reproductive and respiratory syndrome virus (PRRSV), porcine epidemic diarrhea virus (PEDV), pseudorabies virus (PRV), and porcine circovirus 2 (PCV2), as well as an atypical porcine pestivirus (APPV) positive cDNA sample (Table S2). Fifty spleen tissues were collected during 2016 to 2018 from pigs with one or more suspected clinical signs, including pyrexia, huddling, weakness, conjunctivitis, and diarrhea. In addition, national certified reference material of African swine fever virus (ASFV) standard genomic DNA (genotype II) was purchased from the China Animal Health and Epidemiology Center.
LwaCas13a expression bacterial glycerol stock was inoculated LB medium for 2.5 h at 37°C and then induced with 0.5 mmol/L isopropyl-β-D-thiogalactopyranoside (IPTG) for 15 h at 16°C. Cell pellets were harvested and resuspended in lysis buffer (20 mmol/L Tris-HCl, pH 8.0, 500 mmol/L NaCl, 1 mmol/L dithiothreitol [DTT]). After sonication, the supernatant was collected for nickel affinity chromatography. The 5-mL nickel Sepharose column (GE Healthcare, USA) was washed with lysis buffer containing 20 mmol/L imidazole to remove nontarget proteins, followed by the elution of LwaCas13a with lysis buffer containing 100 mmol/L imidazole. The elution fractions were then incubated with SUMO protease (Solarbio, China) for 8 h at 4°C to cleave off the His6-Twinstrep-SUMO tag. After SUMO protease digestion, the proteins were diluted with SP elution buffer (20 mmol/L Tris-HCl, pH 8.0, 1 mmol/L DTT, 5% glycerol) in a volume ratio of 1:1. Diluted proteins were then loaded onto a 1-mL HiTrap SP cation exchange column (GE Healthcare, USA) and eluted over a NaCl gradient from 250 mmol/L to 1 mol/L. Fractions containing LwaCas13a were analyzed by SDS-PAGE, pooled, and stored at −80°C for further study.
CSFV genomic cDNA alignment of Shimen strain and HCLV strain showed a 12-nt insertion (5′-CTTTTTTCTTTT-3′) in HCLV 3′UTR region compared with Shimen 3′UTR region, which was coincident with previous study (33). Candidate crRNAs were designed by individually submitting partial 3′UTR regions of Shimen strain (5′-ACCCUAUUGUAGAUAACACUAAUUUUUUAUUUAUUUA-3′) and HCLV strain (5′-UACACUACUUUUCUUUUUUCUUUUUUAUUUAUU-3′) to CRISPR-RT (Table S3) (34). Candidate crRNAs were in vitro transcribed by the T7 RiboMAX Express large-scale RNA production system (Promega, USA) and purified by NucAway spin columns (Thermo Fisher, USA) according to the manufacturer’s instructions. Double-stranded DNA templates for transcription were prepared by annealing T7 promoter appending forward DNA oligonucleotide with its complementary reverse DNA oligonucleotide in annealing buffer (Beyotime Biotech, China). Each oligonucleotide pair was annealed by preheating the annealing cocktail (nuclease-free water 40 μL, 5× annealing buffer 20 μL, 50 μmol/L forward and reverse oligonucleotide 20 μL each) at 95°C for 2 min and then gradient decreasing 1°C every 90 s from 94 to 25°C. PUC57 plasmids that contained CSFV 3′UTR DNA sequences of different genotypes were used as the templates for PCR amplification with forward primer (5′-TAATACGACTCACTATAGGGG-3′) and reverse primer (5′-TTAGGAAATTACCTTAGTCCAAC-3′) in a 50-μL cocktail containing 2× PrimeSTAR Max Premix (TaKaRa, China) 25 μL, 10 μmol/L forward and reverse primer 2.5 μL each, double-distilled water 18 μL and PUC57 plasmid 2 μL. The PCR steps were as follows: initial denaturation at 98°C for 1 min; 30 cycles at 98°C for 10 s, 45°C for 5 s, and 72°C for 30 s; and a final extension at 72°C for 5 min. PCR products were purified by Cycle Pure kit (Omega Bio-tek, China) and directly used as in vitro transcription templates. CSFV 3′UTR RNAs were then transcribed in vitro by T7 RiboMAX Express large-scale RNA production system (Promega, USA) and purified by MEGAclear transcription clean-up kit (Thermo Fisher, USA) according to the manufacturer’s instructions.
Spleen tissue homogenates were prepared with a grinder in a ratio of 1:5 (tissue weight/phosphate-buffered saline [PBS] volume, mg/μL). Then, 500-μL viral cell cultures or spleen tissue homogenates were freeze-thawed three times and centrifuged at 12,000 × g for 20 min. All procedures were performed at 4°C. Total viral nucleic acids were extracted with MiniBEST Viral RNA/DNA extraction kit version 5.0 (TaKaRa, China). Briefly, 300 μL supernatant was used for extraction procedures according to the manufacturer’s instruction, followed by elution with 30 μL nuclease-free water, and immediately used for cDNA preparation or stored at −80°C for further experiments. cDNAs were prepared with PrimeScript RT Master Mix (TaKaRa, China) according to the manufacturer’s instructions. A 10-μL reverse transcription cocktail contained 5× PrimeScript RT Master Mix (2 μL), 1 μL total viral nucleic acids (less than 500 ng), and 7 μL nuclease-free water. Once reverse transcription was done, cDNAs were stored at −80°C for further experiments.
A single 50-μL cocktail contained 1× reaction buffer (20 mmol/L HEPES, pH 6.8, 60 mmol/L NaCl, 6 mmol/L MgCl2), 50 nmol/liter LwaCas13a, 25 nmol/liter crRNA, 1.6 unit/μL murine RNase inhibitor (New England Biolabs, USA), 125 nmol/liter quenched fluorescent RNA reporter (Thermo Fisher, USA), and various amounts of input nucleic acid target (5 μL). If input nucleic acid target was amplified DNA, the above reaction was modified to include 0.5 μL T7 transcription enzyme mix and 2.5 μL T7 transcription buffer (Promega, USA), as well as 2 μL amplified DNA. LwaCas13a collateral cleavage assay was proceeded for 30 min at 37°C on POLARstar Omega multifunction reader (BIO-GENE Biotech, China) with fluorescent kinetics (excitation/emission [Ex/Em] = 490 nm/520 nm) measured every 3 min.
RT-RAA was performed with basic RT-RAA kit (ZC Bioscience, China) according to the manufacturer’s instructions. A 50-μL RT-RAA cocktail included 41.5 μL rehydration buffer (buffer A), 1 μL each of 20 μmol/L T7 promoter appending forward primer (5′-TAATACGACTCACTATAGGGGGGGAACCCGCCAGTAGGACCCTATTGTAGATA-3′) and reverse primer (5′-GTGGTAACTTGAGGTAGTTTGTACCAGTTCTT-3′), 4 μL RNA template, and 2.5 μL 280 mmol/L MgOAc (buffer B). RT-RAA cocktail was performed at 37°C in a water bath for 30 min and immediately used as amplified DNA for LwaCas13a collateral cleavage assay. CSFV nested PCR (nPCR) was performed according to OIE recommend and previous study with some modifications (35). Primary PCR was performed with outer primer pairs (outer-forward-primer 5′-CAACTGGCTVGTYAAYGC-3′ and outer-reverse-primer 5′-AATGAGTGTAGTGTGGTAAC-3′, V = A or G; Y = C or T) in a 25-μL cocktail, which contained 12.5 μL 2× rTaq mix (TaKaRa, China), 1 μL 10 μmol/L outer forward or reverse primer each, 8 μL double-distilled water, and 2.5 μL 10-fold diluted cDNA. The first round PCR steps were as follows: initial denaturation at 95°C for 1 min; 25 cycles at 94°C for 30 s, 54°C for 30 s, and 72°C for 30 s; and a final extension at 72°C for 10 min. Amplified DNA of first round PCR was then used as the templates for second-round PCR, with inner primer pairs (inner-forward-primer 5′-ATGATGATGVCSCTKATA-3′ and inner-reverse-primer 5′-GTGTGGTAACWTGAGGTAG-3′, V = A or G; Y = C or T, V = A or G; S = C or G; K = T or G; W = A or T) in a 25-μL cocktail, which contained 12.5 μL of 2× rTaq mix, 1.5 μL of 10 μmol/L inner forward or reverse primer, 9 μL of double-distilled water, and 0.5 μL of first round PCR product. The second-round PCR steps were as follows: initial denaturation at 95°C for 3 min; 35 cycles at 94°C for 30 s, 56°C for 30 s, and 72°C for 20 s; and a final extension at 72°C for 10 min. In addition, the primers and conditions of nPCRs for BVDV, APPV, PEDV, PRRSV, PRV, and PCV2, as well as ASFV OIE-recommended PCR, are listed in Table S4.
CSFV antigen ELISA was performed with a commercial classical swine fever virus antigen test kit/serum plus (IDEXX, China). Supernatants of viral cell cultures or tissue homogenates were mixed with detection solution in a volume ratio of 1:1 and incubated at 37°C for 2 h, followed by adding conjugate to each well and incubated at room temperature for another 30 min. Color development was performed with TMB substrate at room temperature for 10 min and stopped by stop solution. The results were calculated according to the manufacturer’s instructions. In brief, optical density at 450 mm (OD450) values of negative-control (N), positive-control (P), and tested samples (S) were measured with POLARstar Omega multifunction reader (BIO-GENE Biotech, China). For a reliable experiment, the N value should be less than 0.250, while the value of (P – N) should more than 0.150. The results were then calculated according to the following rules: a positive result was determined if the value of (S-N) was more than 0.300; a suspected result was determined if the value of (S – N) was between 0.100 and 0.300; and a negative result was determined if the value of (S – N) was less than 0.100.
HUDSON was performed for robust treatment of tested samples according to previous study (18). In brief, supernatants of viral cell cultures or tissue homogenates were mixed with EDTA and TCEP at final concentrations of 1 and 100 mmol/L, respectively, followed by boiling for 5 min to lyse viral particles and inactivate nucleases. The mixture then was directly used for RT-RAA and LwaCas13a collateral cleavage assay. | true | true | true |
PMC9603910 | 36214691 | Xiaowei Yan,Yanni Feng,Yanan Hao,Ruqing Zhong,Yue Jiang,Xiangfang Tang,Dongxin Lu,Hanhan Fang,Manjree Agarwal,Liang Chen,Yong Zhao,Hongfu Zhang | Gut-Testis Axis: Microbiota Prime Metabolome To Increase Sperm Quality in Young Type 2 Diabetes | 10-10-2022 | type 2 diabetes in youth,sperm concentration,sperm motility,spermatogenesis,gut microbiota,DHA,EPA | ABSTRACT Young type 2 diabetes (T2D) affects 15% of the population, with a noted increase in cases, and T2D-related male infertility has become a serious issue in recent years. The current study aimed to explore the improvements of alginate oligosaccharide (AOS)-modified gut microbiota on semen quality in T2D. The T2D was established in young mice of 5 weeks of age with a blood glucose level of 21.2 ± 2.2 mmol/L, while blood glucose was 8.7 ± 1.1 mM in control animals. We discovered that fecal microbiota transplantation (FMT) of AOS-improved microbiota (A10-FMT) significantly decreased blood glucose, while FMT of gut microbiota from control animals (Con-FMT) did not. Sperm concentration and motility were decreased in T2D to 10% to 20% of those in the control group, while A10-FMT brought about a recovery of around 5- to 10-fold. A10-FMT significantly increased small intestinal Allobaculum, while it elevated small intestinal and cecal Lactobacillus in some extent, blood butyric acid and derivatives and eicosapentaenoic acid (EPA), and testicular docosahexaenoic acid (DHA), EPA, and testosterone and its derivatives. Furthermore, A10-FMT improved liver functions and systemic antioxidant environments. Most importantly, A10-FMT promoted spermatogenesis through the improvement in the expression of proteins important for spermatogenesis to increase sperm concentration and motility. The underlying mechanisms may be that A10-FMT increased gut-beneficial microbes Lactobacillus and Allobaculum to elevate blood and/or testicular butyric acid, DHA, EPA, and testosterone to promote spermatogenesis and thus to ameliorate sperm concentration and motility. AOS-improved gut microbes could emerge as attractive candidates to treat T2D-diminished semen quality. IMPORTANCE A10-FMT benefits gut microbiota, liver function, and systemic environment via improvement in blood metabolome, consequently to favor the testicular microenvironment to improve spermatogenesis process and to boost T2D-diminished semen quality. We established that AOS-improved gut microbiota may be used to boost T2D-decreased semen quality and metabolic disease-related male subfertility. | Gut-Testis Axis: Microbiota Prime Metabolome To Increase Sperm Quality in Young Type 2 Diabetes
Young type 2 diabetes (T2D) affects 15% of the population, with a noted increase in cases, and T2D-related male infertility has become a serious issue in recent years. The current study aimed to explore the improvements of alginate oligosaccharide (AOS)-modified gut microbiota on semen quality in T2D. The T2D was established in young mice of 5 weeks of age with a blood glucose level of 21.2 ± 2.2 mmol/L, while blood glucose was 8.7 ± 1.1 mM in control animals. We discovered that fecal microbiota transplantation (FMT) of AOS-improved microbiota (A10-FMT) significantly decreased blood glucose, while FMT of gut microbiota from control animals (Con-FMT) did not. Sperm concentration and motility were decreased in T2D to 10% to 20% of those in the control group, while A10-FMT brought about a recovery of around 5- to 10-fold. A10-FMT significantly increased small intestinal Allobaculum, while it elevated small intestinal and cecal Lactobacillus in some extent, blood butyric acid and derivatives and eicosapentaenoic acid (EPA), and testicular docosahexaenoic acid (DHA), EPA, and testosterone and its derivatives. Furthermore, A10-FMT improved liver functions and systemic antioxidant environments. Most importantly, A10-FMT promoted spermatogenesis through the improvement in the expression of proteins important for spermatogenesis to increase sperm concentration and motility. The underlying mechanisms may be that A10-FMT increased gut-beneficial microbes Lactobacillus and Allobaculum to elevate blood and/or testicular butyric acid, DHA, EPA, and testosterone to promote spermatogenesis and thus to ameliorate sperm concentration and motility. AOS-improved gut microbes could emerge as attractive candidates to treat T2D-diminished semen quality. IMPORTANCE A10-FMT benefits gut microbiota, liver function, and systemic environment via improvement in blood metabolome, consequently to favor the testicular microenvironment to improve spermatogenesis process and to boost T2D-diminished semen quality. We established that AOS-improved gut microbiota may be used to boost T2D-decreased semen quality and metabolic disease-related male subfertility.
Type 2 diabetes (T2D) has long been considered a disease of adulthood; however, there has been a steep rise of this disease in children and adolescents worldwide (1–5), parallel to the increasing rates of obesity. It is reported the incidence of T2D in children <15 years of age in New Zealand has increased progressively at 5% per year over the past 2 decades (2). The reasons for this increase include genetic factors, environmental factors, and a lack of physical activity (1–5). T2D patients are at potentially higher risk for most of the outcomes compared to type 1 diabetes (T1D) patients. For example, 72% of T2D patients display different types of early diabetes-related complications; however, just 32% of T1D patients have similar evidence (3). The most impressive feature of T2D in youth is that it is more commonly diagnosed at lower age and a lower body mass index in boys than in girls (4). Moreover, the majority of youths with T2D present at a mean age of 13.5 years (1–5). Accumulating data from human and animal studies indicate that gut bacteria play fundamental roles in T2D, as there is profound dysbiosis in T2D (6–9). Gut microbiota and hosts have developed a coherent symbiotic relationship where the gut microbiota contribute to many physiological functions, including energy metabolism, metabolic signaling, immune system formation, and gut barrier integrity (9). Gut microbiota benefits humans through the production of short-chain fatty acids (SCFAs) from the fermentation of carbohydrates; moreover, a deficiency in SCFAs is correlated with T2D (8). Bifidobacterium, Bacteroides, Faecalibacterium, Akkermansia, and Roseburia are reported to be negatively associated with T2D; however, Ruminococcus, Fusobacterium, and Blautia are positively associated. Lactobacillus is frequently detected and reported in T2D and has produced inconsistent results among investigations (7, 9). In humans and experimental animal models, Bacteroides has been shown to be a beneficial microbe for glucose metabolism (6). Moreover, microbes have been considered a treatment for T2D (7), especially as dietary intervention can modulate the gut microbiota to improve glucose status in T2D (7). Male infertility is already a common health problem worldwide, and up to one in six couples have infertility issues (10, 11). Diabetes poses an adverse impact on male fertility directly and indirectly at multiple levels, such as impairment to spermatogenesis itself, penile erection, and ejaculation (12–15). Thus, male infertility issues have become more common with the increasing rates of T2D. The negative impacts of diabetes on erectile and ejaculation function, as well as a reduction in semen volume, sperm count, sperm motility, and abnormal sperm morphology, have been reported widely (13). Since T2D-related male infertility is such a serious issue, much effort has been directed to overcome it. The hydroalcoholic extract of Rhus coriaria seeds, resveratrol, metformin, and chitosan-stabilized selenium nanoparticle combination, adiponectin, nesfatin-1, and testosterone have been used in the treatment of T2D-impaired semen quality and related male infertility (16–22). Recently, our group and others have found that gut dysbiosis disrupts spermatogenesis to decrease semen quality and/or male fertility (23–26). Moreover, gut microbiota transplantation (FMT) has been shown to be an effective approach for improving semen quality (23, 24). However, it is not understood whether gut microbiota could effectively ameliorate semen quality in T2D subjects. The current study aimed to explore the beneficial improvement of alginate oligosaccharide (AOS)-modified gut microbiota on semen quality in T2D since it has been found to be effective in ameliorating semen quality in busulfan-treated subjects (23, 24).
Three-week-old male mice were fed with a high-fat diet (HFD) for 2 weeks, after which one dose of streptozotocin (STZ; 85 mg/kg body weight) was injected intraperitoneally (i.p.). After 3 days, blood glucose (21.2 ± 2.2 mmol/L [mM]) was significantly higher in the HFD plus STZ (HS) group than in the control (Con) group (8.7 ± 1.1 mM), which indicated that a T2D model has been successfully established (blood glucose > 11.1 mM) (27, 28). Then, mice in the HS group were divided into three groups, (i) the HS group (STZ+HFD), (ii) the A10-FMT group (HFD+STZ with fecal microbiota transplantation [FMT] from AOS-improved gut microbiota), and (iii) Con-FMT group (HFD+STZ plus FMT from control animal gut microbiota) (see Fig. 1a for study scheme and see Materials and Methods for details). After another 5 weeks, the body weight and blood insulin levels were lower in the HS, A10-FMT, and Con-FMT groups than that in the Con group (see Fig. S1a and b in the supplemental material). Blood glucose was significantly higher in the HS group than that in the Con group; however, it was significantly reduced by A10-FMT, but not by Con-FMT (Fig. 1b), which indicated that A10-FMT treatment improved T2D status. At the same time, blood glycogen was decreased by both A10-FMT and Con-FMT (Fig. 1c). Sperm concentration and motility were significantly diminished in T2D subjects (HS group), while A10-FMT (not Con-FMT) significantly increased sperm motility and concentration (Fig. 1d and e). Gut microbiota has been reported to be disturbed in T2D patients and animal models (6, 9). Small intestine, cecum, and colon content microbiota were determined in the current investigation (Fig. 2; Fig. S1c to n; Fig. S2). In the small intestine, Allobaculum was increased in the A10-FMT group compared to Con and HS groups (Fig. 2a to d), while Desulfovibrio was decreased in the HS, A10-FMT, and Con-FMT groups compared to Con (Fig. 2a to d). In the cecum, Bacteroides were increased in the A10-FMT and Con-FMT groups compared to Con or HS (Fig. S2a to d), while both Coprococcus and Flexispira were decreased in the HS, A10-FMT, and Con-FMT groups compared to Con (Fig. S2a to d); however, Helicobacter was increased in the HS group compared to Con, while it was decreased in the A10-FMT and Con-FMT groups (Fig. S2a to d). On the other hand, Lactobacillus was decreased in the HS, while it was increased in the A10-FMT and Con-FMT groups (Fig. S2a to d), although not significantly. In the colon, Bacteroides and Helicobacter were altered in the same trend as in the cecum (Fig. S2e to h). The function of changed gut microbe genes was enriched by Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis, and eight major signaling pathways were disturbed by HS, while they were reversed in the A10-FMT and/or Con-FMT groups in the small intestine, cecum, and/or colon (Fig. 2e). Interestingly, the energy metabolism pathway was increased in the colon, while it was decreased in the small intestine and cecum in the A10-FMT group specifically, and the lipid metabolism pathway was decreased in the A10-FMT group, while it was increased in the Con-FMT group in the colon (Fig. 2e). However, the glycan biosynthesis and metabolism and metabolism of cofactor and vitamins pathways were decreased in the HS group, while they were increased in both the A10-FMT and Con-FMT groups in the cecum and colon (Fig. 2e). The membrane transport pathway was increased by HS, while it was decreased in both the A10-FMT and Con-FMT groups in the colon (Fig. 2e), and the cell motility pathway was increased by HS, while it was decreased in both the A10-FMT and Con-FMT groups in the cecum (Fig. 2f). Moreover, the carbohydrate metabolism pathway was decreased by HS, while it was increased in the A10-FMT and Con-FMT groups in the small intestine. On the other hand, the carbohydrate metabolism pathway was increased by HS, while it was decreased in the A10-FMT and Con-FMT groups in the colon (Fig. 2e). In total, the data indicated that A10-FMT and Con-FMT may differentially modulate gut microbiota and microbial function to regulate blood metabolites and other functions such as spermatogenesis in type 2 diabetes.
T2D is a metabolic disease. Gut microbiota regulates the blood metabolome, which was confirmed in the current study by liquid chromatography-mass spectrometry (LC-MS) (Fig. S3; Data Set S1). Compared to the Con group, blood triglyceride (TG) was increased in the HS group, which was decreased in the A10-FMT group; however, the data were not significant. Cholesterol was higher in the HS group, while it was decreased in the A10-FMT group (Fig. 3a and b). At the same time, blood lipid molecules were increased in the HS group; however, they were reduced by A10-FMT and/or Con-FMT (Fig. 3c and d), which is consistent with gut microbiota data. Bile acids play vital roles in lipid and other metabolism in the intestine and liver (25). The most common bile acids were increased in the HS and Con-FMT groups, while they were decreased in the A10-FMT group (Fig. 3e to g). It is very interesting to note that butyric acids and derivatives were decreased in the HS group, while they were increased in the A10-FMT group (Fig. 3h to j). The blood n-3 polyunsaturated fatty acids (PUFA) and eicosapentaenoic acid (EPA) were lower in the HS group, while they increased in the A10-FMT group (Fig. 3k). Similarly, blood retinoic acid was in a decreasing trend in the HS group, while it was in an increasing trend in the A10-FMT group; however, the data were not significant (Fig. 3l). It should be noted that in some cases, the data were not significant between the Con and HS groups (Fig. 3j to l), which suggests that the changes observed may not be representative for T2D. The correlation analysis of blood metabolites and sperm concentration/motility showed that blood glucose and cholesterol were negatively correlated with sperm concentration and sperm motility, while EPA and butyric acid were positively correlated with sperm concentration and motility (Fig. 3m), which suggests that the blood metabolome and semen quality parameters are correlated. The liver is the major metabolic organ that is responsible for most of the metabolites in the blood. Liver function may be disturbed by HFD and STZ treatment (HS), while A10-FMT reverses the damage, as shown by the histopathology of the liver samples in different groups (Fig. S4a). The transcriptome sequencing (RNA-seq) analysis of liver samples showed that HS disrupted gene expression in the liver, while it was recovered by A10-FMT and/or Con-FMT (Fig. S4b to d). Functional enrichment analysis found that liver lipid metabolism function was upset by HS (Fig. S4c). Expression of the lipid metabolism-related genes (such as NR1H3, Acaa1a, Acaa2, Acox1, Acox2, etc.) was decreased by HS, while it increased was by A10-FMT and/or Con-FMT (Fig. S4c). Nuclear receptor subfamily 1 group H member 3 (NR1H3), a key regulator of lipid homeostasis, has been found to be involved in lipid deposition in pig (29). The protein level of NR1H3 was increased in the HS group, while it was decreased in the A10-FMT group but not in the Con-FMT group (Fig. 4a and f). Blood bile acids were significantly higher in the HS and Con-FMT groups, which was consistent with the protein levels of the bile acid-produced enzyme CYP7A1 (Fig. 4b and f) and bile acid receptor NR1H4 (Fig. 4c and f), which were higher in the HS and Con-FMT groups, while they were reduced in the A10-FMT group. Retinoic acids play vital roles in lipid metabolism in the liver (29). The protein levels of retinoic synthesis protein DHRS9 and binding protein RBP4 were decreased by HS, while they were increased by A10-FMT (Fig. 4d, e, and g). All data together indicated that liver lipid metabolism function was disrupted by the HS group and recovered by A10-MFT, which matched the blood metabolism data. Moreover, HS caused apoptosis in liver cells, which further suggested HS damaged liver function, while it was recovered by A10-FMT (Fig. 4h and i). At the same time, A10-FMT increased systemic antioxidant capability through the total antioxidant capability (T-AOC) of the blood and GPX1 in liver (Fig. 4h to j); however, the data were not significant between Con and HS (Fig. 4h to j), which suggests that the changes observed may not be representative for T2D.
Since A10-FMT improved liver function and the blood metabolome, next, we set out to explore the effects of A10-FMT on the testicular metabolome (Fig. 5; Fig. S5; Data Set S2). n-3 PUFAs such as docosahexaenoic acid (DHA) and EPA play vital roles in spermatogenesis (30, 31). It is interesting to note that A10-FMT, but not Con-FMT, increased DHA and EPA and their derivatives in the testes (Fig. 5a to f). Moreover, many other unsaturated acids were also increased by A10-FMT (Fig. 5g to i) in the testes. However, there is no significant difference between the Con and HS groups for some of the compounds (Fig. 5a, c, f, i, k, and n to p), which suggests that the changes observed may not be representative for T2D. Another group of altered metabolites in the testes was retinoids. There is an increasing trend for the retinoids by A10-FMT compared to HS (Fig. 5j to p). HS significantly decreased retinyl ester and retinal, while A10-FMT significantly increased them (Fig. 5i and m). It is known that retinoids are crucial for spermatogenesis (10, 11). Male steroid hormones, especially testosterone, play vital roles in spermatogenesis. HS treatment significantly decreased testosterone and epitestosterone, while A10-FMT led to a recovery (Fig. 6a and b). HS treatment significantly decreased androsterone, while it was in an increasing trend in A10-FMT (Fig. 6c), which, however, may not be relevant to T2D, given that the data were not significant between the Con and HS groups. The levels of hormone production proteins in the testes, which were decreased by HS, were increased by A10-FMT (Fig. 6d and e). At the same time, the changed testicular metabolites were well correlated with each other in different comparisons, including HS/Con, A10-FMT/HS, and Con-FMT/HS, respectively (Fig. S5g to i). Furthermore, correlation analysis of testicular metabolites and sperm concentration and motility found that there was a good relationship between testicular metabolites and semen quality parameters (sperm concentration and motility [Fig. 6f]). DHA, EPA, and testosterone were positively correlated with sperm concentration and/or sperm motility (Fig. 6f). At the same time, there was a good relationship between blood metabolites and testicular metabolites (Fig. 6g), where some of the metabolites were positively correlated, while others were negatively correlated.
The metabolome data in blood and testicular samples suggested that HS upset the systemic environment and testicular microenvironment to disrupt spermatogenesis and, in turn, reduce sperm concentration and motility; meanwhile, A10-FMT improved the systemic and testicular environment to recover spermatogenesis, sperm concentration, and motility. A10-FMT recovered spermatogenesis by increasing important spermatogenesis proteins such as VASA (germ cell marker), SYCP3 (meiosis marker), PGK2 (sperm motility and fertility protein), ODF1 (component of filamentous structure), PIWIL1 (germ line integrity), and TP1 (spermatid-specific protein) in testicular samples (Fig. 7a to f). However, there is no significant difference between the Con and HS groups for PIWIL1 and TP1; the SYCP3 data were not significant for the A10-FMT or Con-FMT groups compared to the HS group. The Sertoli cell marker, SOX9, was not changed in all the groups (Fig. 7g). The data indicate that A10-FMT improved spermatogenesis and thus increased sperm concentration and motility.
As more youths develop T2D, the risk of complications is high; therefore, many studies are focusing on this disease. It has been reported that T2D in youths is more frequently diagnosed at a lower age and body mass index in boys than girls (4). Furthermore, the mean age of T2D in youths is 13.5 years, which is around the initiation of puberty (1–5), a time when the male reproductive system begins rapid development. This may be the main reason for male infertility (diminished semen quality) in T2D during later reproductive age since hyperglycemia in T2D causes reproductive system dysfunction (13). In the current study, in order to mimic youth T2D patients, 3-week-old male mice were fed with HFD and injected with STZ to induce T2D, and during puberty (5 weeks of age), the mice developed T2D (32, 33). Sperm concentration and motility were decreased, while blood glucose was increased in T2D mice when adult (10 weeks of age), which is consistent with clinical findings (15). However, FMT from AOS-improved gut microbiota (A10-FMT) decreased blood glucose and increased sperm concentration and motility, while Con-FMT did not. The data suggest that AOS-improved gut microbiota contains beneficial microbes that can produce or help with the production of beneficial compounds to ameliorate the systemic environment and testicular microenvironment to improve spermatogenesis. In the current investigation, sperm motility and concentration were increased by A10-FMT compared to T2D (HS group); however, the pregnancy rate was not determined. Gut microbiota plays a crucial role in T2D development (6–9). Bacteroides and Bifidobacterium are beneficial genera that are most frequently reported in studies of T2D (6). In the current study, we found that Bacteroides was increased by A10-FMT and Con-FMT in both the cecum and colon. Data suggest that FMT may be helpful in glucose metabolism, as it has been previously reported that Bacteroides beneficially supports glucose metabolism (6). Actually, blood glycogen was decreased by both A10-FMT and Con-FMT. Helicobacter, one of the gastric-pathogenic bacteria (34, 35), was increased in the cecum and colon in T2D mice, while it was reduced by A10-FMT and Con-FMT in the current investigation. Although Lactobacillus may be considered a beneficial bacterial in T2D, the data were not constant (6). Lactobacillus was decreased in T2D animals, while it was strongly increased by A10-FMT, although not significantly in the current investigation. Another interesting microbe, Allobaculum, an SCFA-producing microbe (35–39), was significantly increased by A10-FMT, while it was increased by Con-FMT (not significantly) compared to the HS group in the small intestine in the current study. At the same time, we found that the blood SCFA, butyric acid, and its derivatives were decreased in T2D and increased by A10-FMT (Con-FMT also increased them to some extent); moreover, blood butyric acid and sperm concentration and motility were positively correlated with each other. It is known that SCFA deficiency is associated with T2D, and dietary fiber can selectively promote gut bacteria to produce SCFAs to alleviate T2D (8). Our findings further confirmed that gut microbiota selectively promoted by dietary components may produce beneficial compounds to attenuate T2D and even improve sperm concentration and motility in T2D, as SCFAs act as signaling molecules in humans (8, 40, 41). T2D is a metabolic disease (3–5). The gut microbiota is involved in metabolism to produce specific molecules in the blood to benefit systemic health (8), and metabolism plays a vital role in spermatogenesis (42). A10-FMT improved gut microbiota by increasing beneficial microbiota to then benefit blood metabolism. A10-FMT reduced blood TG, cholesterol, glucose, and glycogen, while it increased butyric acid and derivatives and EPA. Furthermore, blood glucose and cholesterol were negatively correlated with sperm concentration and motility, while EPA and butyric acid were positively correlated with sperm concentration and motility. Furthermore, A10-FMT improved T2D-impaired liver function to support systemic metabolism. The testicular microenvironment is important for spermatogenesis (42). A10-FMT improved the testicular microenvironment through the significant increase in testicular levels of DHA, EPA, and testosterone and its derivatives. n-3 PUFAs, especially DHA, are crucial for spermatogenesis and male reproductive functions (30, 31, 43). Moreover, testicular metabolites and blood metabolites were well correlated. All the data indicated that A10-FMT improved the systemic environment and testicular microenvironment to promote spermatogenesis to increase sperm concentration and motility. Indeed, it has previously been shown that A10-FMT improves spermatogenesis impaired by busulfan through amelioration of the important proteins in spermatogenesis (23, 24). Although many studies have tried different approaches to ameliorate semen quality in T2D subjects, it was previously unknown whether gut microbiota could improve spermatogenesis to increase semen quality in youths with T2D. For the first time, we established that AOS-improved gut microbiota ameliorated sperm concentration and motility and glucose status in youths with T2D. In summary, this novel investigation in youthful mice with T2D demonstrated that A10-FMT ameliorated gut microbes Lactobacillus and Allobaculum to elevate blood butyric acid and EPA and testicular DHA, EPA, and steroid hormones such as testosterone to promote spermatogenesis, thus increasing sperm concentration and sperm motility. The findings from the current study shed new light on the mechanism of action of beneficial gut microbiota in youths with T2D who have potentially high risks of male infertility. Our findings strongly suggest the need to explore microbiota therapy (FMT) to reduce the elevated rates of male infertility in youths with T2D early in their disease course. Clinical studies in youths with T2D are warranted to elucidate the roles of beneficial microbiota in the improvement of semen quality and male fertility.
All animal procedures used in this study were approved by the Animal Care and Use Committee of the Institute of Animal Sciences of Chinese Academy of Agricultural Sciences (IAS2020-106). The male mice were used in the current study due to its focus on male reproductive health. The mice were maintained in specific-pathogen-free environment under a light/dark cycle of 12:12 h at a temperature of 23°C and humidity of 50% to 70%; they had free access to food (chow diet) and water (14). (i) Experiment I: mouse small intestine microbiota collection. Three-week-old ICR male mice were dosed with double-distilled water (ddH2O) as the control or AOS per 10 mg/kg body weight (BW) via oral gavage (0.1 mL/mouse/day) (23, 24). The AOS dosing solution was freshly prepared every day and was delivered every morning for 3 weeks. There were two groups (30 mice/treatment), (i) control (ddH2O), and (ii) A10 (AOS per 10 mg/kg BW). After 3 weeks treatment, the animals were maintained on regular diet for 3 more days (no treatment). Then, the mice were humanely euthanized to collect small intestinal luminal content (microbiota). (ii) Experiment II: high-fat diet (HFD) and streptozotocin (STZ) treatment and microbiota transplantation (FMT). The small intestine luminal content (microbiota) from each group was pooled and homogenized, diluted 1:1 in 20% sterile glycerol (saline), and frozen (23, 24, 44, 45). Before inoculation, fecal samples were diluted in sterile saline to a working concentration of 0.05 g/mL and filtered through a 70-μm cell strainer. STZ was from Sigma (catalog no. S0130). Three-week-old ICR male mice were used in the current investigation. There were four treatment groups (30 mice/treatment), including (i) Con (control; regular diet plus saline); (ii) HFD+STZ (HS) (HFD from 3 to 10 weeks of age; at 5 weeks old, one dose of STZ at 85 mg/kg body weight after preliminary screening) (32, 33); (iii) Con-FMT (HFD from 3 to 10 weeks of age; at 5 weeks old, one dose of STZ at 85 mg/kg body weight; gut microbiota transplantation [FMT] from control mice [experiment I] from 5 to 7 weeks of age); and (iv) A10-FMT (HFD from 3 to 10 weeks of age; at 5-week-old one dose of STZ at 85 mg/kg body weight; FMT from AOS 10-mg/kg dosed mice [experiment I] from 5 to 7 weeks of age). HFD started from 3 weeks old (continued until the end of the experiment), and one dose of STZ was injected i.p. at 5 weeks old (2 weeks after HFD feeding) (32, 33). Then, the mice received oral FMT inoculations (0.1 mL) once daily for 2 weeks (5 weeks of age to 7 weeks of age) (23, 24). Then, the mice were regularly maintained (on respective diets) for another 3 weeks (10 weeks of age). Then, the mice were humanely euthanized to collect samples for different analyses (Fig. 1a, study scheme).
Spermatozoa motility was assessed using a computer-assisted sperm assay (CASA) method according to World Health Organization guidelines (46). After euthanasia, spermatozoa were collected from the cauda epididymis of mice and suspended in Dulbecco’s modified Eagle medium (DMEM)-F-12 medium with 10% FBS and incubated at 37.5°C for 30 min; samples were then placed in a prewarmed counting chamber. The Microptic sperm class analyzer (CASA system) was used in this investigation. It was equipped with a 20-fold objective, a camera adaptor (Eclipse E200; Nikon, Japan), and a camera (acA780-75gc; Basler, Germany), and it was operated by an SCA sperm class analyzer (Microptic S.L.). The classification of sperm motility was as follows: grade A linear velocity, >22 μm s−1; grade B linear velocity, <22 μm s−1; curvilinear velocity, >5 μm s−1; grade C curvilinear velocity, <5 μm s−1; and grade D, immotile spermatozoa. The spermatozoa motility data represented only grade A and grade B since only these two grades are considered to be functional.
The extracted murine caudal epididymides were placed in RPMI medium and finely chopped, and then, eosin Y (1%) was added for staining as described previously (46). Spermatozoan abnormalities were then viewed using an optical microscope and were classified into head or tail morphological abnormalities, two heads, two tails, blunt hooks, and short tails. The examinations were repeated three times, and 500 spermatozoa per animal were scored.
Briefly, total RNA was isolated using TRIzol reagent (Invitrogen) and purified using a PureLink RNA minikit (catalog no. 12183018A; Life Technologies) following the manufacturers’ protocol (46). Total RNA samples were first treated with DNase I to degrade any possible DNA contamination. Then, the mRNA was enriched using oligo(dT) magnetic beads. Mixed with the fragmentation buffer, the mRNA was broken into short fragments (about 200 bp), after which the first strand of cDNA was synthesized using a random-hexamer primer. Buffer, deoxynucleoside triphosphate (dNTP), RNase H, and DNA polymerase I were added to synthesize the second strand. The double-stranded cDNA was purified with magnetic beads. Subsequently, 3′-end single-nucleotide adenine (A) addition was performed. Finally, sequencing adaptors were ligated to the fragments. The fragments were enriched by PCR amplification. During the quality control (QC) step, an Agilent 2100 bioanalyzer and ABI StepOnePlus real-time PCR system were used to qualify and quantify the sample library. The library products were prepared for sequencing in an Illumina HiSeq 2500 system. The reads were mapped to reference genes using SOAPaligner (v.2.20) with a maximum of two nucleotide mismatches allowed at the parameters of -m 0 -x 1000 -s 40 -l 35 -v 3 -r 2. The read number of each gene was transformed into reads per kilobases per million reads (RPKMs), and then differentially expressed genes were identified using the DEGseq package and the MA plot-based method with random sampling model (MARS) method. The threshold was set as false-discovery rate (FDR) of ≤0.001 and an absolute value of log2 ratio ≥1 to judge the significance of the difference in gene expression. Then, the data were analyzed by gene ontology (GO) and KEGG enrichment.
(i) DNA extraction. Total genomic DNA of small intestine, cecum, and colon digesta was isolated using an E.Z.N.A. stool DNA kit (Omega Bio-tek, Inc., USA) following the manufacturer’s instructions. DNA quantity and quality were analyzed using NanoDrop 2000 (Thermo Scientific, USA) and 1% agarose gel. Ten samples per group were determined. (ii) Library preparation and sequencing. The V3-V4 region of the 16S rRNA gene was amplified using the primers MPRK341F (5′-ACTCCTACGGGAGGCAGCAG-3′) and MPRK806R (5′-GGACTACHVGGGTWTCTAAT-3′) with barcoding. The PCRs (total, 30 μL) included 15 μL Phusion high-fidelity PCR master mix (New England Biolabs), 0.2 mM primers, and 10 ng DNA. The thermal cycle was carried out with an initial denaturation at 98°C, followed by 30 cycles of 98°C for 10 s, 50°C for 30 s, 72°C for 30 s, and a final extension at 72°C for 5 min. PCR products were purified using a GeneJet gel extraction kit (Thermo Scientific, USA). The sequencing libraries were constructed with NEBNext Ultra DNA library prep kit for Illumina (NEB, USA) following the manufacturer’s instructions, and index codes were added. Then, the library was sequenced on the Illumina HiSeq 2500 platform, and 300-bp paired-end reads were generated at the Novo gene. The paired-end reads were merged using FLASH (v.1.2.71). The quality of the tags was controlled in QIIME (v.1.7.02); meanwhile, all chimeras were removed. The “core set” of the Greengenes database was used for classification, and sequences with >97% similarity were assigned to the same operational taxonomic units (OTUs). (iii) Analysis of sequencing data. Operational taxonomic unit abundance information was normalized using a standard sequence number corresponding to the sample with the least sequences. The alpha diversity index was calculated with QIIME (v.1.7.0). The UniFrac distance was obtained using QIIME (b1.7.0), and principal-coordinate analysis (PCoA) was performed using R software (v.2.15.3). The linear discriminate analysis effect size (LEfSe) was performed to determine differences in abundance; the threshold LDA score was 4.0. GraphPad Prism7 software was used to produce the graphs.
Plasma samples were collected and immediately stored at −80°C. Before LC-MS/MS analysis, the samples were thawed on ice and processed to remove proteins. Testis samples were collected, and the same amount of tissue from each mouse testis was used to isolate the metabolites using CH3OH to H2O (vol/vol) ratio of 4:1. Then samples were detected by Acquity UPLC and AB Sciex TripleTOF 5600 (LC/MS) as reported previously (23, 46). Ten samples per group were analyzed for plasma or testis samples. The HPLC conditions employed an Acquity UPLC ethylene-bridged hybrid (BEH) C18 column (100 mm by 2.1 mm, 1.7 μm), solvent A (aqueous solution with 0.1% [vol/vol] formic acid), and solvent B (acetonitrile with 0.1% [vol/vol] formic acid) with a gradient program. The flow rate was 0.4 mL/min, and the injection volume was 5 μL. Progenesis QI v.2.3 (Nonlinear Dynamics, Newcastle, UK) was implemented to normalize the peaks. Then, the Human Metabolome Database (HMDB), LIPID MAPs (v.2.3), and METLIN software were used to qualify the data. Moreover, the data were processed with SIMCA software (v.14.0; Umetrics, Umeå, Sweden) following pathway enrichment analysis using the KEGG database (https://www.genome.jp/kegg/pathway.html).
Blood insulin was determined by the kit from Beijing Solarbio Science & Technology Co., Ltd. (Beijing, People’s Republic of China; catalog no. SEKM0141). Blood alanine aminotransferase (ALT; catalog no. C009-2-1), aspartate transaminase (AST; catalog no. C010-2-1), TG (catalog no. A110-1-1), and T-AOC (catalog no. A015-2-1) were determined by the kits from Nanjing Jiancheng Bioengineering Institute (Nanjing, People’s Republic of China) (47). All procedures were followed from the manufacturers’ instructions.
Testicular tissues were fixed in 10% neutral buffered formalin, paraffin embedded, cut into 5-μm sections. and subsequently stained with hematoxylin and eosin (H&E) for histopathological analysis (46).
Western blotting of proteins was carried out as previously reported (23, 46). Briefly, testicular tissue samples were lysed in radioimmunoprecipitation assay (RIPA) buffer containing the protease inhibitor cocktail from Sangong Biotech, Ltd. (Shanghai, China). Protein concentration was determined using a bicinchoninic acid (BCA) kit (Beyotime Institute of Biotechnology, Shanghai, China). Goat anti-actin was used as a loading control. The information for primary antibodies (Abs) is listed in Table S1 in the supplemental material. Secondary donkey anti-goat Ab (catalog no. A0181) was purchased from Beyotime Institute of Biotechnology, and goat anti-rabbit (catalog no. A24531) Abs were bought from Novex by Life Technologies (USA). Fifty micrograms of total protein per sample was loaded onto 10% SDS-polyacrylamide electrophoresis gels. The gels were transferred to a polyvinylidene fluoride (PVDF) membrane at 300 mA for 2.5 h at 4°C. The membranes were then blocked with 5% bovine serum albumin (BSA) for 1 h at room temperature (RT), followed by three washes with 0.1% Tween 20 in TBS (TBST). The membranes were incubated with primary Abs diluted 1:500 in TBST with 1% BSA overnight at 4°C. After three washes with TBST, the blots were incubated with the horseradish peroxidase (HRP)-labeled secondary goat anti-rabbit or donkey anti-goat Ab, respectively, for 1 h at RT. After three washes, the blots were imaged. The bands were quantified using ImageJ software. The intensity of the specific protein band was normalized to actin first, and then the data were normalized to the control. The experiment was repeated >6 times.
The methodology for immunofluorescence staining of testicular samples is reported in our recent publications (23, 46). Sections of testicular tissue (5 μm) were prepared and subjected to antigen retrieval and immunostaining as previously described. Briefly, sections were first blocked with normal goat serum in phosphate-buffered saline (PBS), followed by incubation with primary Abs (1:100 in PBS-0.5% Triton X-100; Bioss Co. Ltd. Beijing, People’s Republic of China) (Table S1) at 4°C overnight. After a brief wash, sections were incubated with an Alexa 546-labeled goat anti-rabbit secondary Ab (1:100 in PBS; Molecular Probes, Eugene, OR, USA) at RT for 30 min and then counterstained with 4′,6-diamidino-2-phenylindole (DAPI). The stained sections were examined using a Leica laser scanning confocal microscope (Leica TCS SP5 II; Germany). Ten animal samples from each treatment group were analyzed. Positively stained cells were counted. A minimum of 1,000 cells were counted for each sample of each experiment. The data were then normalized to the control.
Data were analyzed using SPSS statistical software (IBM Co., NY, USA) with one-way analysis of variance (ANOVA) with least significant difference (LSD) multiple comparison. The data were shown as the mean ± standard error of the mean (SEM). Statistical significance was based on P values of <0.05.
All animal procedures used in this study were approved by the Animal Care and Use Committee of the Institute of Animal Sciences of Chinese Academy of Agricultural Sciences.
Liver RNA-seq raw data were deposited in NCBI’s Gene Expression Omnibus under accession number GSE179098. The microbiota raw sequencing data generated in this study have been uploaded to the NCBI SRA database with the accession number PRJNA742204 (small intestine), PRJNA742202 (cecum), and PRJNA742203 (colon). | true | true | true |
PMC9604055 | Dinah Farhanah Jamal,Quratul Ain Rozaimee,Nadila Haryani Osman,Atikah Mohd Sukor,Marjanu Hikmah Elias,Nor Aripin Shamaan,Srijit Das,Nazefah Abdul Hamid | Human Papillomavirus 16 E2 as an Apoptosis-Inducing Protein for Cancer Treatment: A Systematic Review | 19-10-2022 | apoptosis,cell death,E2 protein,human papillomavirus,HPV 16 | Human papillomavirus type 16 (HPV-16) is a well-known etiological factor for cervical and oropharyngeal cancers. The E2 protein, the product of an early-transcribed gene in HPV–16, is postulated to cause the death of cancerous cells via p53-dependent and p53-independent pathways. The main aim of the present systematic review was to study the HPV 16-E2 protein as an apoptosis-inducer agent. A thorough search of MEDLINE/PubMed, Science Direct, Scopus, and EBSCOhost databases was conducted for relevant studies on HPV AND apoptosis OR cell death where HPV 16-E2 was involved. The search identified 967 publications. Eleven records dated from 1 January 1997 to 16 February 2022 were found to meet the inclusion criteria and were eligible for data extraction and inclusion. All studies concluded that HPV 16-E2 was able to induce cell death in transfected cells. E2 proteins from the high-risk HPV–16 were able to induce apoptosis through different apoptotic pathways depending on the location of the expressed gene. However, the mechanism was still unclear, and further studies are warranted. | Human Papillomavirus 16 E2 as an Apoptosis-Inducing Protein for Cancer Treatment: A Systematic Review
Human papillomavirus type 16 (HPV-16) is a well-known etiological factor for cervical and oropharyngeal cancers. The E2 protein, the product of an early-transcribed gene in HPV–16, is postulated to cause the death of cancerous cells via p53-dependent and p53-independent pathways. The main aim of the present systematic review was to study the HPV 16-E2 protein as an apoptosis-inducer agent. A thorough search of MEDLINE/PubMed, Science Direct, Scopus, and EBSCOhost databases was conducted for relevant studies on HPV AND apoptosis OR cell death where HPV 16-E2 was involved. The search identified 967 publications. Eleven records dated from 1 January 1997 to 16 February 2022 were found to meet the inclusion criteria and were eligible for data extraction and inclusion. All studies concluded that HPV 16-E2 was able to induce cell death in transfected cells. E2 proteins from the high-risk HPV–16 were able to induce apoptosis through different apoptotic pathways depending on the location of the expressed gene. However, the mechanism was still unclear, and further studies are warranted.
Cervical cancer has emerged as the fourth-most common cancer affecting women with a 6.5% incidence and 7.7% mortality reported worldwide [1]. The major risk of cervical cancer includes human papillomavirus (HPV) infection, smoking, age, and the use of contraceptives [2]. Human papillomavirus (HPV) is classified into low- and high-risk infectious agents depending on their ability to cause malignancies in normal cells. High-risk HPV types 16 (HPV–16) and 18 (HPV-18) are reported to be responsible for HPV-positive tumors with a 70% global prevalence for cervical cancer [3]. It has been reported that in women aged 15 years and older, HPV-16 is the most common genotype, followed by HPV types 52, 31, and 18 [4]. The HPV genome consists of early and late gene regions where gene arrangements influence the sequence of gene expression and controls the infection mechanism of HPV. Each of the early genes designated as E1, E2, E4, E5, E6, and E7 has specific roles in gene expression and infection [5]. HPV can remain in the epithelial basal layer and cause latent infection in the basal cells. Moreover, encapsulated HPV infects the stratified squamous epithelium cells that are exposed through microlesions. Consequently, viral promoters are activated, resulting in expression of E1 and E2 [6]. E1 and E2 genes encode for the protein complex that leads to E1 attachment to the origin of replication, which helps in regulating the transcription of the subsequent viral genome [7,8]. The E4 protein blocks cell apoptosis by binding with cytokeratin and it is believed to play a role in virus release in order to initiate infection of nearby cells. The E5 protein helps promote cell proliferation. The binding of E6 and E7 to p53 and retinoblastoma (pRb), respectively, leading to DNA replication, even in non-dividing cells. The late gene region mainly encodes major and minor structural capsid proteins, such as L1 and L2. capsid proteins enable the complete genome to exit the infected cells as an encapsulated infectious virus [9]. Targeting the major capsid L1 to activate humoral antibody responses has become the major principle in developing the HPV vaccine, which is now commercially available [10]. The bivalent Cervarix® vaccine consists of virus-like particles (VLP) L1 from HPV-16 and HPV-18, while the quadrivalent Gardasil® vaccine consists of VLP L1 from HPV types (6, 11, 16, and 18), thus providing wider protection against HPV [10]. Gardasil 9, a nine-valent vaccine, has been approved by FDA in 2014 and offers protection against HPV-6, 11, 16, 18, 31, 33, 45, 52, and 56 [11]. The 9-valent HPV vaccine provides more protection against HPV compared to the quadrivalent HPV vaccine, is safe, and its cost-effectiveness favours its use in adolescent females [11]. The E2 gene encodes the E2 protein, which is required for other proteins to function, and regulates transcription through interactions with the E1 viral protein in DNA replication initiation [9]. It is involved in cellular gene expression, growth inhibition, and apoptosis [12]. In HPV-transformed malignant cells, the chromosomal integration between HPV and the host cell disrupts the E2 open reading frame, leading to the deduction that the absence of E2 protein expression is a vital step for the deregulation of the cell cycle, and the initiation of carcinoma transformation [13,14,15]. Nevertheless, the HPV-16 E2 protein has the ability to induce apoptosis through more than one pathway, which is through p53-dependent and p53-independent pathways [16,17,18,19]. Both HPV-transformed and non-transformed cells have p53-dependent pathways, whereas HPV-transformed cells only have p53-independent pathways [17]. There are variable opinions among researchers regarding the pathways and proteins involved in E2 apoptotic functions. E2 proteins from high-risk HPV types 16 and 18, which are associated with cancer of cervix, were reported to induce apoptosis [16]. In addition to controlling transcription and viral DNA replication, E2 proteins physically interacted with cellular proteins to have an impact on the biology of the host cell [16]. The involvement of the E2 protein in apoptosis occurred via the p53-dependent and p53-independent pathway, receptor-signaling pathway, and mitochondria-dependent pathway [16,17,18,19,20,21]. Published studies showed that E2 induces apoptosis indirectly, via its effects on the expression of E6 and E7, and directly, via its interaction with p53 [17]. Another mechanism is that there is the binding of E2 to the viral genome, and this can only happen in HPV-transformed cells [17]. E2 possesses multi-functional properties and has a role on the normal viral life cycle of keratinocytes [18]. The viral DNA is amplified and encapsidated by viral structural proteins as the host cell matures until mature virus particles are shed with squames from the epithelial surface [18]. Hence, a systematic review was conducted to evaluate the HPV-16 E2 protein as an apoptosis-inducing agent in both HPV-transformed and non-HPV-transformed cells. This systematic review may pave the way to finding novel methods of using the E2 protein as a potential treatment strategy for HPV infection-related cancer.
A comprehensive search strategy was conducted to assemble the available literature on the apoptosis-inducing function of human papillomavirus type 16 E2 viral proteins. Four well-known research databases were chosen, and search strategies were designed according to Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines 2020.
A thorough literature search was conducted using four databases: MEDLINE/PubMed, EBSCOhost, ScienceDirect, and Scopus were searched from 1 January 1997 to 16 February 2022. Two investigators (D.F.J and N.A.H) conducted the title and abstract search using keywords chosen for the widest search coverage of any possible studies on the HPV-16 E2 protein but simultaneously retaining the specific objective of the study. The searches were not restricted to any publication year or language. The following keywords were used for the literature search; “Human papillomavirus 16 E2”, “HPV 16 E2”, 1 or 2 or 3, “Apopto*”, “programmed cell death”, 5 or 6, “cancer”, “malignancy”, “malignant neoplasm”, “neoplasia”, “neoplasm”, 7 or 8 or 9 or 10 or 11, 4 and 7, and 13. The asterisk sign indicates the “wild-card” search technique, where all terms that have the same root word as stated were included in the search process. Step 1. All results obtained from each search database were imported into a separate individual group in one combined Endnote (X7.0.7; Thomson Reuters) library file, where duplicate results were identified and deleted. Step 2. Abstract and articles’ keywords were reviewed for eligibility by two of the investigators (D.F.J and N.A.H). Titles and abstracts were thoroughly screened, and irrelevant articles were removed. The following exclusion criteria were applied during screening as a guide to remove some of the results from the study selection: (i) study on types other than human papillomavirus type 16; (ii) study proteins on other than E2 viral protein; (iii) company, clinical trial summary reports, in vivo studies, or in silico studies; and (iv) studies published in languages other than English. Any studies deemed as ambiguous underwent a second screening where full papers were retrieved and carefully re-evaluated based on the inclusion criteria. Step 3. The following inclusion criteria were applied during screening to finalize the study selection: (i) investigation on high-risk HPV type 16; (ii) involvement of the E2 protein as the subject of interest; (iii) investigation of the viral protein’s function in programmed cell death or apoptosis; (iv) test conducted on cervical cancer cells; (v) original research paper (not reviews or proceedings); and (vi) in vitro studies.
The references from the finalized articles were screened and examined for any possible related studies. The full-text articles were then retrieved and reviewed to ascertain if they met the inclusion criteria.
Screened papers that made the cut were reviewed, and information such as first author, year of publication, study outcome, type of cell lines, type of vector or co-factor expressing the E2 protein, and methods to determine apoptosis and levels of observed apoptosis were extracted. No restrictions were made in study design or publication date. A data evaluation form was used as a guide to extract the required information, as shown Table 1.
As this review focuses on E2’s apoptotic function, all studies must include an apoptosis assay measurement or cell viability test in the methodology. Once confirmed, further evaluation was conducted on the level of apoptosis and the methods used to measure the level of apoptosis.
This review aimed to qualitatively investigate the effect of the HPV-16 E2 protein with respect to apoptosis functions. Thus, meta-analysis could not be attempted due to the study’s exploratory design.
A total of 391 records were identified from MEDLINE/PubMed, and 36 from ScienceDirect, 745 from EBSCOHOST, and 165 records from Scopus were selected, resulting in a total of 967 records. Using EndNote X7.0.1 (Thomson Reuters), 64 duplicate results were removed (Figure 1). Titles and abstracts from the 903 recorded articles were reviewed for their validity. A total of 878 articles were excluded for various reasons, e.g., using proteins other than E2 protein as a cell death causative agent, focusing on different cancerous cells such as head and neck cancers, and different types of high-risk papillomavirus such as HPV type 18 or low-risk papillomavirus such as HPV-11 and HPV-6. Finally, a total of 11 records dating from 1 January 1997 to 16 February 2022 were found to meet the inclusion criteria. All 11 records are experimental study designs involved in testing the HPV–16 E2 protein in vitro. There was only one in vivo study where the combination of radiation treatment and fusion protein E2 was carried out in mice [20]. A total of four studies were conducted in the United Kingdom (UK) [22,23,24,25] and China [19,20,21,26], respectively, two in India [27,28], and another one in Mexico [29]. The two studies in India [27,28] and China [19,20] were carried out by the same team of researchers, presenting the possibility of reporting bias. The stable or transient expressions of E2 protein in cell lines were the main procedures used to study the protein’s function in vitro. HPV-16 and HPV-18 transformed cervical cancer cell lines, and SiHa and HeLa were the most popular cancer cell lines in these studies. SiHa was used in eleven experiments [19,20,21,22,23,24,25,26,27,28,29], while HeLa was used in five experiments [20,23,24,25,27]. C33A was the most common HPV-negative squamous cell carcinoma used as a comparison; it was used in six of the selected studies [19,20,21,24,26,29]. COS-7, the SV40-transformed monkey fibroblast cell line, has been used in two experiments [23,24] to test for E2 protein expression. MCF-7, an HPV-negative human breast adenocarcinoma, has been used in three experiments [23,25,27]. CasKi and HEK293 have been used in three [23,25,29] and two experiments [25,27] each. Other cell lines such as BMK-16/myc, 778, 877, 915, W12, and B16 were used once in separate experiments [23,24,25,29]. The apoptotic effect was seen in all HPV-transformed cell lines, such as HeLa [20,23,24,25,27], SiHa [19,20,21,22,23,24,25,26,27,28,29], Saos-2 [25], and some of the HPV-negative cell lines such as C33A [19,20,21,24,26,29], MCF-7 [23,25,27], and NIH3T3 [24,25]. Table 2 shows the summary of these cell lines.
The increased or decreased level of cell death was recorded based on the percentage of the cell’s population. We discovered that the most common method for determining apoptosis was by flow cytometry; nine studies used the method [19,20,21,22,23,25,26,27,28]. Six of the studies used either propidium iodide (PI) or Annexin V staining [19,21,22,23,25,27], while another two of the studies used a combination of both [20,26], and one study did not mention the details [28]. Six studies used fluorescence microscopy to inspect cell morphology for apoptosis analysis [19,22,23,24,25,29]. Three of them stated the use of Hoechst 33258 dye to detect apoptotic cells [19,23,24]. Four studies applied the Terminal deoxynucleotidyl transferase dUTP nicked end labelling (TUNEL) assay [19,20,25,29]. Six studies have combined more than one method to determine and measure the level of cell death [19,20,22,23,25,29]. All studies performed an empirical comparison to differentiate cell death caused by apoptosis or necrotic response either based on morphology assessment or by using flow cytometry readings [19,20,21,22,23,24,25,26,27,28,29]. Alongside cell death analysis, three studies used the MTT (3–(4,5–Dimethylthiazol–2–yl)–2,5 Diphenyltetrazolium Bromide) or water-soluble tetrazolium (WST–1) assay (another form of MTT assay) to measure cell viabilities in HPV-16 E2 contained vector transfection [20,22,29]. The cell viability percentage stated was to support the E2 protein apoptotic effect on the cell population. For data collection purposes, the stated value in the records was the average of at least three independent experiments, either for the apoptosis or cell viability test. The t-test using mean and standard deviation was the statistical analysis employed in this study.
Vectors such as recombinant adenoviral plasmid have been used extensively to transfer the E2 gene expression into the cell line of interest. The HPV-16 E2 expression plasmid was produced by cloning the E2 open reading frame gene sequence into EcoR1 restriction sites in pWEB, together with the downstream cytomegalovirus (CMV) promoter to produce pWEB-E2 [23]. This eukaryotic expression vector (pWEB-E2) was used in three separate studies [23,24,25]. Three studies have been identified to use the same pcDNA 3.1 expression system to clone the HPV–16 E2 gene [19,21,26]. The polymerase chain reaction (PCR) was used to amplify the HPV–16 E2 gene before cloning the plasmid of interest. The cytomegalovirus (CMV) promoter and the green fluorescent protein (GFP) were incorporated to improve transcription and assist in protein identification. Two studies co-transfected pCMX-GFP3, which expressed the GFP protein to identify the transfected cells and allow the assessment of cellular morphology [23,24]. Two studies used plasmid pCB6 + p53 for wild-type p53 [22,24], while pCB6 + p53173L represented a mutant p53 plasmid for investigating the p53 interaction [22,24]. In order to control the E2 gene expression in the clone vector, there was a heavy-metal inducible metallothionein promoter used, and the transfection took place upon the induction of cadmium [22]. Other studies used combination treatments by using co-factors such as steroid hormones and radiation to enhance the E2 apoptosis action [20,24]. Among the 11 studies, 6 studies [19,21,26,27,28,29] used Lipofectamine. Only one study [25] used FuGENE 6, Tfx-20, and Tfx-50 as transfection reagents to transfer the expression vector into the cell line of interest. One study [27] applied conventional calcium–phosphate precipitation methods. The summary is shown in Table 3. Information summary of the study’s outcome, type of cell line, plasmid construct, methods to measure apoptosis, and the level of apoptosis is shown in Table 1.
Regarding the present systematic review, we found 11 studies [19,20,21,22,23,24,25,26,27,28,29] on the apoptotic function of HPV-16 E2 that fits our parameters. The re-introduction of HPV-16 E2 protein into the cells was shown to reduce cell growth and promote cell death in serum-starved cells [22]. The E2 protein was shown to increase E6 and E7 mRNA levels. The E7 binds to Rb and releases the free E2F-1 protein, which is believed to induce apoptosis via a p53-dependent pathway. Admittedly, this systematic review did not include studies on HPV types other than HPV-16. The current systematic review showed that only the E2 from HPV-16 E2 is capable of inducing apoptosis, as reported by Blachon et al. [16] and Parish et al. [17]. It has been suggested that the difference in E2 protein’s ability to induce cell death between low and high–risk HPV types was caused by the intracellular localization of the protein itself rather than its proteomic properties [16,17]. The E2 protein from low-risk HPV type 6 and HPV type 11 remained in the nucleus, while the high-risk E2 protein was present in both the nucleus and cytoplasm. The accumulation of E2 in the cytoplasm activates the caspase 8 cascade, which then leads to apoptosis [16]. There was a report that highlighted the difference between low and high-risk E2 in terms of p53 interaction [17]. The inability of the low-risk E2 protein to bind with p53 fails to induce apoptosis. As for the high-risk E2, E2-p53 binding is required in both HPV-transformed and non-transformed cells in order to induce apoptosis. However, it was suggested that there is a second pathway involved specifically in HPV-transformed cells, where the E2 protein is capable of inducing apoptosis through its interaction with the viral genome E6 or E7 [16]. In another study, the HPV-16 E2 was expressed in a variety of cell lines and showed that the protein was able to induce apoptosis in both HPV-transformed cells and in at least two non-HPV transformed cell lines, C33A and COS–7 cells [24]. The author concluded that apoptosis was p53-dependent and did not require DNA–E2 protein binding. Additionally, the same author in the following year reported that steroid hormones, estrogen, and progesterone could be used to enhance the apoptotic effect induced by the HPV-16 E2 protein [30]. The interaction between E2 protein and p53 was also supported by experiments carried out by Brown et al. [31]. The apoptosis event was measured in both HeLa cells and Saos-2 cells; it was observed that the E2–p53 interaction caused the downregulation of HPV DNA. There was also a study that succeeded in inducing cell death in non-transfected bystander cells using the modified herpes simplex virus VP22 fusion protein with E2; pVP22-E2 [23]. The expressed VP22-E2 protein was shown to cause apoptosis in infected tumors in a concentration-dependent manner. This finding suggests the possibility of using VP22–E2 as a post-surgical treatment to eliminate remaining tumor cells. However, the further elucidation of this possibility needs to be validated. It has been reported on the apoptotic role of the E2 protein in radiation treatment in vivo and in vitro [20]. The HPV-16 E2 gene, which was cloned into the novel oncolytic adenoviral M5, was demonstrated to increase the cytotoxic effect on transfected cells in vitro. In vivo, the M5 appeared to enhance the efficacy of radiation treatment in the mice tumor model. Although the exact mechanism between M5-E2 and radiation was not explained, the author suggested a connection to the death-receptor-signalling pathway as caspase 8 activity increased [20]. Despite prior evidence of HPV-16 E2 inducing a high apoptosis rate, a study indicated that there was prolonged stable expression of E2 proteins in human keratinocyte cell lines (HaCaT) that survived the apoptotic effect, providing first insight into E2 protein’s role in malignant transformation after infection [18]. Researchers continued to investigate other possible protein interactions in the apoptosis pathway, and it has been reported that the hyperactivation of caspase-8 and caspase-3 caused by the overexpression of high–risk HPV was mediated by cellular–FLICE (FADD-like IL–1β-converting enzyme)–inhibitory protein (c-FLIP) [19]. Additionally, one study showed that the globular heads of the C1q receptor (gC1qR) play a role in inducing apoptosis in C33A and SiHa cervical cancer cells [21]. gC1qR is a protein embedded in the outer mitochondrial membrane, which is known to mediate many biological responses, including the initiation of apoptosis [32]. The study found that in human cervical squamous carcinoma samples showed significantly lower expression of the HPV-16 E2 and gC1qR genes than non-cancerous cervix samples. In addition, when gC1qR small-interfering RNA (siRNA) was added, the gC1qR gene expression, mitochondrial malfunction, and cellular death were greatly elevated in C33a and SiHa cells that had been transfected with a vector encoding HPV-16 E2. These findings support a mechanism whereby gC1qR exerts a significant influence in HPV-16 E2-induced human cervical squamous carcinoma cell apoptosis via a mitochondria-dependent pathway [21]. C33A cells are HPV-negative cell lines that express mutated p53 [33]. The p53-dependent apoptosis action was deduced when no cell death was observed in HPV-16 E2 protein-infected cells unless it was co–expressed with p53 [24]. COS-7 did not undergo E2-induced apoptosis, thus making it an ideal choice to compare the tagged protein expression in vitro [23]. Saos-2 is a p53-null cell line and it is more likely to be used when the study investigates the effect of p53 in the presence of the E2 protein [25]. In terms of techniques used to measure apoptosis, PI/Annexin V staining via flow cytometry was able to provide a quantitative analysis of the proliferation and condition of the transfected cells where the exact population of cells’ condition could be quantified [34]. Apoptosis analysis via the TUNEL assay generally produces lower readings as it cannot detect late apoptotic cells that have detached. On the other hand, internucleosomal DNA fragmentation detected by the TUNEL assay can be caused by non-apoptotic cells, including necrotic cell death, cells undergoing DNA repair, and cells damaged by other forces, which could lead to the overestimation of reading [35]. Therefore, PI/Annexin V flow cytometry is more accurate compared to the other apoptosis assays employed in the studies. PI/Annexin V staining is capable of differentiating the cell lysates into early and late apoptosis phases, allowing the investigator to obtain data that are more accurate. In comparison of five studies that used both SiHa (HPV type 16) and HeLa (HPV type 18), no significant difference was found between the levels of apoptosis caused by different HPV strains [20,23,24,25,27].
In addition to controlling viral gene expression, the HPV E2 protein is necessary for viral replication. The current systematic review showed that HPV-16 E2 induced apoptosis in both HPV-transformed and non-HPV-transformed cells. The tumor-suppressor gene p53 is responsible for the human genome stability and has the potential to repair DNA. The E2 gene interaction with p53 is vital in non-HPV-transformed cells but not in HPV-transformed cells. The involvement of E2 gene in the p53-dependent and p53-independent apoptotic pathways makes the E2 gene a suitable candidate for the therapeutic targeted gene not only for cervical cancer but also potentially for other cancer types as well. Nevertheless, further studies on larger sample size are warranted to elucidate the apoptosis mechanism and the apoptotic effect of E2 protein. | true | true | true |
|
PMC9604060 | Andranik Ivanov,Daniele Mattei,Kathrin Radscheit,Anne-Claire Compagnion,Jan Patrick Pett,Hanspeter Herzel,Rosa Chiara Paolicelli,Monika Piwecka,Urs Meyer,Dieter Beule | Analyses of circRNA Expression throughout the Light-Dark Cycle Reveal a Strong Regulation of Cdr1as, Associated with Light Entrainment in the SCN | 15-10-2022 | circular RNA (circRNA),Cdr1as,circadian rhythm,suprachiasmatic nucleus,SCN,non-coding RNA,Cyrano,light entrainment,miR-7,light-dark,wake-sleep | Circular RNAs (circRNAs) are a large class of relatively stable RNA molecules that are highly expressed in animal brains. Many circRNAs have been associated with CNS disorders accompanied by an aberrant wake-sleep cycle. However, the regulation of circRNAs in brain homeostasis over daily light-dark (LD) cycles has not been characterized. Here, we aim to quantify the daily expression changes of circRNAs in physiological conditions in healthy adult animals. Using newly generated and public RNA-Seq data, we monitored circRNA expression throughout the 12:12 h LD cycle in various mouse brain regions. We identified that Cdr1as, a conserved circRNA that regulates synaptic transmission, is highly expressed in the suprachiasmatic nucleus (SCN), the master circadian pacemaker. Despite its high stability, Cdr1as has a very dynamic expression in the SCN throughout the LD cycle, as well as a significant regulation in the hippocampus following the entry into the dark phase. Computational integration of different public datasets predicted that Cdr1as is important for regulating light entrainment in the SCN. We hypothesize that the expression changes of Cdr1as in the SCN, particularly during the dark phase, are associated with light-induced phase shifts. Importantly, our work revises the current beliefs about natural circRNA stability and suggests that the time component must be considered when studying circRNA regulation. | Analyses of circRNA Expression throughout the Light-Dark Cycle Reveal a Strong Regulation of Cdr1as, Associated with Light Entrainment in the SCN
Circular RNAs (circRNAs) are a large class of relatively stable RNA molecules that are highly expressed in animal brains. Many circRNAs have been associated with CNS disorders accompanied by an aberrant wake-sleep cycle. However, the regulation of circRNAs in brain homeostasis over daily light-dark (LD) cycles has not been characterized. Here, we aim to quantify the daily expression changes of circRNAs in physiological conditions in healthy adult animals. Using newly generated and public RNA-Seq data, we monitored circRNA expression throughout the 12:12 h LD cycle in various mouse brain regions. We identified that Cdr1as, a conserved circRNA that regulates synaptic transmission, is highly expressed in the suprachiasmatic nucleus (SCN), the master circadian pacemaker. Despite its high stability, Cdr1as has a very dynamic expression in the SCN throughout the LD cycle, as well as a significant regulation in the hippocampus following the entry into the dark phase. Computational integration of different public datasets predicted that Cdr1as is important for regulating light entrainment in the SCN. We hypothesize that the expression changes of Cdr1as in the SCN, particularly during the dark phase, are associated with light-induced phase shifts. Importantly, our work revises the current beliefs about natural circRNA stability and suggests that the time component must be considered when studying circRNA regulation.
Circadian rhythms are physiological, molecular, and behavioral changes that follow a 24-h cycle. Most multicellular organisms take advantage of environmental factors, such as light and temperature, to fine-tune their daily behavior and metabolism to create a coherent circadian system. In mammals, the suprachiasmatic nucleus (SCN) is the central circadian pacemaker [1,2] that is entrained in light-dark (LD) cycles [3]. From the retina, the light signal is transmitted to the SCN, which then synchronizes clocks in other tissues [4]. The endogenous mammalian circadian clock consists of transcription-translation feedback loops [5] that are accompanied by global changes in transcriptional and post-transcriptional regulation [6,7], including RNA synthesis and degradation, alternative splicing, and editing [8,9,10,11]. As of now, it is clear that many RNA species, both protein-coding and non-coding, play an essential role in maintaining the circadian rhythm [6,7,12,13]. However, circular RNAs (circRNAs), an alternative form of RNA splicing products, dependent on temperature and RNA editing [14,15,16,17], have not been well studied throughout daily activities and circadian rhythm. CircRNAs are a large class of highly stable RNA molecules that consist of multiple or single exons [18,19,20]. They have conserved biogenesis and are expressed in animals and plants [14,16,18,19,21]. CircRNAs are primarily localized to the cytoplasm and have various functions, such as sequestering miRNAs, modulation of RNA stability, and multiple interactions with RNA-binding proteins [16,18,22,23]. A subset of circRNAs with retained introns (so-called exon-intron circRNAs) reside in the nucleus and regulate mRNA transcription [24]. Certain circRNAs can be translated into proteins [25,26] or released from cells in extracellular vesicles [27,28]. In animals, circRNAs are highly expressed in the brain, and a subset of circRNAs is enriched at the synaptic terminals [15,29]. Due to their high stability and synaptic enrichment, these molecules are thought to be involved in intracellular information transport, long-term memory formation, and memory consolidation [30]. Moreover, the deregulation of circRNAs has been implicated in neurodegenerative, psychiatric, and neurodevelopmental disorders [31,32,33,34,35]. Many of these brain conditions occur in parallel, or arguably due to aberrant wake/sleep cycles [36,37], which is why it is necessary to study circRNAs in the context of the circadian rhythm. Finally, due to their relatively stable nature and tissue-specific expression pattern, circRNAs have been recognized as biomarkers in different pathologies [38], particularly in many cancer types [39], where circRNAs are systematically deregulated [40]. Given the recent observations that cancer metastases can accelerate during the rest (sleep) phase [41], it becomes evident that the regulation of circRNAs should be analyzed in both the wake and sleep phases. Here, we studied circRNA expression in different mouse brain regions during 12:12 h LD cycles and predicted Cdr1as as an essential regulatory molecule that impacts the light entrainment in the SCN. Cdr1as is highly expressed in the brain, particularly in glutamatergic neurons [31]. It has an unusually high number of miR-7 binding sites [19,22]. MiR-7 is a potent, brain-enriched miRNA [42,43] that has been implicated in the pathogenesis of different brain diseases [44,45,46,47,48,49,50,51,52]. Cdr1as and miR-7 regulate neuronal activity through a network with two other non-coding RNAs, miR-671 and Cyrano [53]. The loss of Cdr1as in mice results in the deregulation of excitatory synaptic transmission and the upregulation of immediate early genes [31]. Cdr1as deletion in mice also leads to vision defects [54]. Our findings explain previously observed gene expression changes and phenotypes in Cdr1as mutant mice and add another piece to the puzzle of Cdr1as functions.
We quantified the circRNA expression levels in the SCN by systematically analyzing the total RNA-Seq data from Pembroke et al. [7]. The authors monitored six equidistant time points in nine replicates during a 12:12 h LD cycle. We identified a total of 1691 expressed circular RNAs in the SCN. The number of consistently detectable circRNAs was highest in the middle of the dark phase, with a substantial drop at its end (325, 485, 455, 450, 523, and 290 circRNAs at ZT2, ZT6, ZT10, ZT14, ZT18, and ZT22, respectively; cf. Figure 1A). Cdr1as circRNA showed the highest expression of all (Figure 1B, Supplementary Table S1). It was expressed one order of magnitude more than any other circRNA in the SCN (Figure 1B). Cdr1as contributed the most to the overall differences in circular reads during the whole 12:12h LD cycle (Figure 1C,D) and was the only highly expressed circRNA with significant regulation in SCN over the LD cycle (e.g., ZT14 vs. ZT10: DESeq2 adj. p-value 1.3 × 10−19). A rather unexpected observation is that this stable circRNA is downregulated more than twofold in four hours (Figure 1D) and again upregulated, exhibiting twin-peak expression with peaks at the beginning and the end of the dark cycle (Figure 1D). Interestingly, Pembroke et al. [7] described a cluster of 766 genes with exact twin-peak expression as a synaptic module regulating light-induced phase shifts in the SCN. Inspecting the Cdr1 locus in the UCSC genome browser showed all reads mapping to the Cdr1 locus map within the circRNA boundaries rather than the primary transcript (Figure 1E). Thus, all reads mapping to the Cdr1 gene could be attributed to the circular RNA, allowing a better Cdr1as expression quantification (Supplementary Figure S1) and putting Cdr1as into the top 5 highest (sorted by FPKM) expressed RNAs in the synaptic twin-peak module described by Pembroke et al. (Figure 1F). Variability between replicates differs substantially between time points. It is highest near the light-dark transitions (Supplementary Figure S1, Figure 1D), which might hint at quick regulatory processes. Interestingly, the long non-coding RNA Cyrano, known to serve as a negative regulator of miR-7 in the brain [53], showed the same twin-peak expression pattern as Cdr1as (Supplementary Figure S1, Figure 1F).
Piwecka et al. [31] showed that loss of Cdr1as results in the upregulation of certain circadian genes (Per1, Sik1, Klf10, Npas4) and upregulation of immediate early genes, such as Fos, Egr1, Klf4, Jun, etc. This group of immediate early genes is activated and rapidly transcribed in the SCN upon photic stimuli [55]. Recently, the Takahashi lab measured genome-wide mRNA changes in the SCN after a short period of light exposure (e.g., 30 min) [56]. This early light-induced gene set is significantly enriched in genes deregulated upon Cdr1as KO in different brain regions (Figure 2A,B). Further, Xu et al. [56] identified that the Npas4 transcription factor is an essential regulator of circadian behavior and transcriptional response to light in the SCN [56]. In the polyA+ RNA-Seq of the Cdr1as KO mice [31], NPas4 is significantly upregulated in three (hippocampus, cerebellum, olfactory bulb) out of the four sequenced brain regions. Moreover, in the hippocampus (of Cdr1as KO mice), Npas4 is the 11th and the cerebellum’s 7th most significantly deregulated gene. To further study the effect of light on Cdr1as regulation, we analyzed the expression of twin-peak cluster genes (defined by Pembroke et al.) in the SCN during constant darkness, using data from Cheng et al. [57]. The study provides RNA-Seq at four circadian time points (CT = 0, 6, 12, 18) in murine SCN with five replicates. Before harvesting, the mice were released from the standard 12:12 h LD cycle into complete darkness for two days (dark-dark or DD cycle). We selected the top 50 highest expressed genes from the twin-peak module and monitored their expression in LD and DD cycles. Due to the polyA enrichment protocol used by Cheng et al., Cdr1as circRNA itself is not detectable in this data. However, none of the 49 other highly expressed genes changed their expression significantly during the DD cycle (Figure 2C,D). Our observations hold when the analysis is carried out using the top 100 or top 200 highest expressed genes (sorted by FPKM) from the twin-peak module (Supplementary Figure S2). This result suggests that the overall expression of the twin-peak cluster and likely Cdr1as in murine SCN highly depends on light exposure. Light is communicated to the SCN from the retina by glutamatergic neurotransmission from the retinohypothalamic tract. Given the Cdr1as enrichment specifically in glutamatergic neurons [31] and high expression of miR-7 in the SCN [58,59], we suggest that Cdr1as may be an important regulator of glutamatergic synaptic transmission during light-induced phase shifts through the delivery of miR-7 to its early response and twin-peak target genes. The overlap of miR-7 targets (444 in total from the Targetscan database [60]) and twin-peak cluster (766) identified 37 genes (Supplementary Table S2) that could be regulated by Cdr1as/miR-7 complex in the context of light-induced phase-shifts. Pembroke et al. [7] reported that the twin-peak cluster is enriched in synaptic transmission and calcium signaling. As expected, its subset miR-7 targets are also significantly enriched in these biological pathways (Supplementary Table S2). Thus, it is conceivable that the Cdr1as:miR-7 complex plays a role in regulating photic inputs transmitted into the SCN (Figure 2E). Among miR-7 target genes from the twin-peak cluster that can be regulated by Cdr1as:miR-7 in the SCN, we can find ionotropic glutamate receptor complex (Grin2a and Grin2b), calcium signaling genes (via Prkcb, Stim2, Hpcal4, and Cacng7), and genes involved in the regulation of the synaptic structure and its activity (Snca, Grin2a, Shank2, Shisa7, Atp2b2, Grin2b, Bcr, and Gabra1) (Figure 2E). The first response to light is likely regulated via Fos, Klf4, and Nr4a3 (Figure 2E).
To further characterize the circadian expression of circRNAs, we sequenced total RNA from two additional brain regions: the hippocampus and frontal cortex. Mice were dissected at six equidistant time points throughout the LD phases. While the SCN acts as a circadian master pacemaker, other brain regions also display oscillatory capacity [61]. The hippocampus is particularly interesting, as it is essential for sleep-dependent memory consolidation [62]. It is interesting to study circRNA regulation in this context, as circRNAs are highly expressed in the hippocampus [15] and have a much longer lifetime than linear RNAs. Therefore, they may be involved in memory consolidation and the transport of information between different cell types [30]. We also investigated the frontal cortex, which is associated with many psychiatric and neurodegenerative disorders. Using three replicates at six time points (SE 150 bp, average sequencing depth 46 mil.), we identified 5505 and 4790 distinct circRNAs expressed in the hippocampus and frontal lobe (Supplementary Table S3). Our analysis of the frontal cortex identified 1328 of 1770 (75%) circbase [63] circRNAs, and in the hippocampus, 1297 of 1676 circbase circRNAs. In the hippocampus, along with a general reduction of exonic reads (Figure 3A) circRNA expression decreased by 20% during the transition from the dark to the light phase (Figure 3B). However, such a change in circularized RNA was most probably due to the difference in the expression of the linear host transcripts (Supplementary Figure S3). The expression of Cdr1as in the hippocampus was downregulated by 30% during the transition from the light to the dark phase and again upregulated during the dark phase. In the frontal cortex, Cdr1as expression did not change significantly during the LD cycle (Figure 3C). Resembling changes in exonic rates (Figure 3A), the overall circRNA expression in the cortex increased by 30% in the wake phase and was downregulated again just before the light phase started (Figure 3B). Similar to the hippocampus, such a fluctuation of circularization was due to a general change in the expression of the linear host gene (Supplementary Figure S3).
Microglia are brain-resident macrophages, and their inflammatory responses are controlled by the intrinsic circadian clock [64]. Microglia also exhibit morphological differences between wake and sleep, and disruption of the clock system of these cells may result in impaired behavior and contribute to sleep disturbance [65]. The regulation of circRNAs in microglia is of particular interest, as these cells are in close contact with synapses and, like neuronal cells, express genes with long introns - a common source of circRNAs [14,20]. In addition to producing circRNAs, microglia can phagocytose circRNAs exported by neurons or astrocytes to the extracellular space. We sequenced freshly isolated microglia from the mouse hippocampus and frontal cortex during the light and dark phases: ZT4 and ZT16 (per time-point 16 replicates, SE 150bp, average sequencing depth: 42 mil.). Our analyses identified 1234 and 1553 distinct circRNAs in the cortex and hippocampus, respectively (Supplementary Table S4). Cdr1as was one of the highest expressed circRNAs in microglia (Figure 3D). Despite an overall increase in circularization during the light phase (Figure 3E), the only circRNA that significantly changed its expression in both the cortex and hippocampus was Cdr1as (Figure 3F). In Cortex, Cdr1as upregulated during the light phase by ~2.6-fold (DESeq2: adj. p-value = 1.6 × 10−7) and in the hippocampus by 1.8 fold (DESeq2: adj. p-value = 0.002).
Our study monitored circRNA expression in the mouse frontal cortex, hippocampus, and suprachiasmatic nucleus in a 12:12 h LD cycle. We characterize Cdr1as circRNA as a novel gene associated with light entrainment in the SCN. We found that Cdr1as and Cyrano, another non-coding RNA involved in miR-7 regulation in the brain, feature oscillatory patterns throughout the LD cycle that replicate the expression pattern of the twin-peak genes. Light-induced regulation of Cdr1as was also observed in the hippocampus, but to a lesser extent compared to SCN. Although there is no evidence that photic impulse directly regulates Cdr1as expression, we show that genes upregulated upon Cdr1as deletion are highly enriched in the light-induced immediate response pathway. Moreover, we demonstrate that the strong regulation of the synaptic twin-peak cluster during the 12 h of the dark cycle (wake phase for mice) occurs only after withdrawal from the light cycle. Pembroke et al. [7] reported that the twin-peak module is highly enriched with genes involved in synaptic transmission, long-term potentiation, calcium signaling, and gated channel activity. On the other hand, Piwecka et al. [31] demonstrated that the loss of Cdr1as circRNA in mice results in aberrant excitatory synaptic transmission. In the retina, deletion of Cdr1as circRNA results in increased beta-wave amplitude of the photopic electrophysiological response and reduced vision contrast sensitivity [54]. Together, these observations suggest that Cdr1as plays an important role in communicating light from the retina to SCN, most likely through the regulation of glutamatergic neurotransmission. Additionally, our results shed more light on the deregulation of certain clock genes in the brain of Cdr1as deficient mice, as reported previously [31]. Loss of Cdr1as also caused the downregulation of a miRNA family (miR-96/miR-182/miR-183), specifically in the cortical region [31]. Only recently has it been discovered that these miRNAs are involved in the modulation of the circadian rhythm [12]. Since two other members of the regulatory network described by Kleaveland et al. [53], miR-7 and Cyrano, are also expressed in the SCN, the central circadian pacemaker emerges as a perfect system for studying the interplay of these non-coding molecules. We believe that further experiments exploring the exact pathway by which Cdr1as influences the activity of SCN neurons should take into account the oscillatory pattern of Cdr1as along with miR-7 targets from the twin-peak module. One surprising observation in the SCN is that Cdr1as, a highly stable circular RNA, is downregulated by more than two-fold during 4 h (from ZT14 to ZT18, Supplementary Figure S1). One possible mechanism may be the cleavage of circRNA by miR-671 [66]. In addition, Cdr1as turnover may occur due to structure-mediated RNA decay by Upf1 and G3bp1 [67]. Both genes are highly expressed in the SCN [7]. We also observed regulation of this abundant circRNA expression throughout the LD cycle in the hippocampus. It is generally accepted that circular RNAs are naturally long-lived in physiological conditions and change their expression significantly only over longer processes (e.g., development, neuronal maturation, or aging) and in pathological conditions (cancer and neurodegeneration). On the contrary, our observations point toward the dynamic nature of circRNA Cdr1as in two brain regions. It is interesting to think about the implications of this finding for cancer. Cdr1as is deregulated in multiple tumor types [68,69,70,71,72,73,74,75], including breast cancer, where it promotes the metastatic phenotype [76]. In addition, the inhibition of Cdr1as increases the sensitivity of drug-resistant breast cancer cells [77]. It was recently shown that the generation of circulating tumor cells in breast cancer does not occur continuously but during the sleep phase [41]. Thus, the characterization of Cdr1as expression patterns over the wake-sleep cycle might bring new insights into the metastasis of cancer cells. Contemplating the importance of these findings for neurodegenerative disorders, the interaction of Cdr1as/miR-7 with alpha-synuclein (Snca) (twin-peak gene) is particularly interesting. Abnormal expression of Snca and its aggregation are critical in the pathophysiology of Parkinson’s disease (PD) [78]. Moreover, in PD patients, miR-7 is significantly downregulated in the brain regions that undergo neurodegeneration during the course of the disease [44]. Due to its binding ability to Snca 3′UTR in vivo and subsequent regulation of protein translation, miR-7 is becoming widely appreciated as an important therapeutic target for PD [44,45,51,52]. Sleep disorders are a common feature of patients in the early stages of the disease [79]. Thus, understanding the strong regulation of Cdr1as in the central circadian pacemaker may provide a new paradigm for studying the early onset of PD. It is important to highlight that the present study demonstrates oscillations in circRNA expression patterns in physiological conditions in adult animals. As research is developing around circRNA dynamics in brain development and specific disorders, it will be highly relevant to determine how the latter are affected in disease from a circadian perspective. In preclinical studies, mice are most often sacrificed for downstream molecular analyses during their sleep phase, while in clinical studies, the time of collection of samples from human patients and controls is rarely taken into consideration as a biologically relevant variable. Altered sleep and circadian patterns are, however, hallmarks of psychiatric and neurodegenerative diseases [80,81]. This study provides an example of how modeling circadian variability, both in coding and non-coding RNAs like Cdr1as, can widen our understanding of transcriptional dynamics in these conditions. Finally, circRNAs are actively screened as biomarkers in many diseases [38,39]. Our findings suggest that their predictivity may be time sensitive, making it important to carry out these experiments in a time-controlled manner.
CircRNAs have not only a developmental stage- but also daytime-dependent expression. Specifically, circular RNA Cdr1as is regulated throughout light-dark cycles in the SCN, which may influence the organism’s adaptation to quick daily changes.
Nine- to ten-week-old male C57BL6/N mice (Charles River Laboratories, Sulzfeld, Germany) were used throughout the study. We allocated 3-5 animals per cage in individually ventilated cages (IVCs). The animal vivarium was a specific pathogen-free (SPF) holding room that was temperature- and humidity-controlled (21 ± 3 °C, 50 ± 10%). Animals used for the microglial isolation were kept under a reversed light-dark cycle (lights off 09:00 AM–09.00 PM), while the animals used for whole hippocampal and frontal lobe tissue sequencing were kept on a normal light-dark cycle (lights on 6 AM–6 PM). All animals had ad libitum access to the same food (Kliba 3436, Kaiseraugst, Switzerland) and water throughout the study. All procedures described in the present study had been previously approved by the Cantonal Veterinarian’s Office of Zurich, and all efforts were made to minimize the number of animals used and their suffering.
Brain tissue dissociation and microglia cell isolation were performed according to a protocol recently optimized and published by us [82]. The protocol was carried out at 4 °C to avoid cell activation during the isolation procedure. Briefly, the animals were deeply anesthetized with an overdose of Nembutal (Abbott Laboratories, North Chicago, IL, USA) and transcardially perfused with 15 mL ice-cold, calcium- and magnesium-free Dulbecco’s phosphate-buffered saline (DPBS, pH 7.3–7.4). The brains were quickly removed and washed with ice-cold DPBS, after which the hippocampi and frontal cortices were dissected on a cooled petri dish and placed in an ice-cold Hibernate-A medium. Mechanical dissociation at 4 °C was carried out on ice. The tissue was dissociated in 1.5 mL Hibernate-A medium in a 1 mL Dounce homogenizer with a loose pestle. The homogenized tissue was then sieved through a 70 μm cell strainer. The homogenates were pelleted at 450× g for 6 min at 4 °C. The supernatants were removed, and the pellets were re-suspended with a P1000 micropipette, applying a pipette-tip cut-off. 500 microliters of freshly prepared isotonic percoll solution was then added to each sample (final volume: 2 mL) and mixed well. Percoll was rendered isotonic by mixing 1 part of 10× calcium- and magnesium-free DPBS (pH 7.3–7.4) with 9-parts of percoll. Importantly, the pH of percoll was adjusted to 7.3–7.4 with 5 molar hydrochloric acid before starting the isolation procedure. The percoll solution was mixed properly with the cell suspension, after which 2 mL of DPBS was gently layered on top of it with a pipette boy set at the slowest speed, creating two separate layers. The samples were centrifuged for 10 min at 3000× g. The centrifugation resulted in an upper layer consisting of DPBS and a lower layer consisting of percoll. The two layers were separated by a disk of myelin and debris, while the cells were located at the bottom of the tube. The layers were aspirated, leaving about 500 μL. The cells were then washed once in DPBS and pelleted by centrifuging them at 460× g for 10 min at 4 °C. This pellet consists of total brain cells, including microglial cells.
Microglia cells were isolated via magnetic-activated cell sorting (MACS) using mouse anti-CD11b magnetic microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany), according to the manufacturer’s instructions, with some modifications. The MACS buffer used consisted of 2% bovine serum albumin (BSA) diluted in DPBS from a 7.5% cell culture-grade BSA stock (Thermo Fisher Scientific Inc., Waltham, MA, USA). Total hippocampal/frontal cortex cell pellets after percoll separation (see above) were re-suspended in 90 μL MACS buffer and 10 μL anti-mouse-CD11b magnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany). The cells were then incubated for 15 min at 4 °C. Cells were washed with 1 mL MACS buffer and pelleted at 300 rcf for 5 min at 4 °C. The cells were then passed through an MS MACS column (Miltenyi Biotec, Bergisch Gladbach, Germany) attached to a magnet. After washing the columns three times with MACS buffer, microglia were flushed from the column with 1 mL of MACS buffer and pelleted at 300 rcf for 5 min at 4 °C. Cell pellets were then snap-frozen in liquid nitrogen and stored at −80 °C.
For the total RNA-seq of whole brain tissue, RNA was extracted from adult male mice total hippocampi and frontal lobes. Briefly, deeply anesthetized adult male mice were intracardially perfused with ice-cold Dulbecco’s phosphate-buffered saline (DPBS) to remove blood. Hippocampi and frontal lobes were dissected immediately afterward in a pre-cooled sterile Petri dish on ice. The brain regions were immediately transferred to an RNAse-free Eppendorf tube, snap-frozen in liquid nitrogen, and stored at −80 °C until RNA extraction.
Total RNA from freshly isolated microglia and from hippocampal and frontal lobe tissue was extracted via phenol/chloroform extraction using the SPLIT-RNA extraction kit (Lexogen, product code: 008.48), according to the manufacturer’s instructions. The RNA was treated with Turbo DNase I (Ambion, product code AM1907) to remove traces of genomic DNA. Following DNase I treatment, the RNA was stored at −80 °C until library preparation.
Before library preparation, the integrity of each sample was assessed on an Agilent TapeStation system 4150 using RNA screen tape (Agilent Technologies Inc., Santa Clara, CA, USA). In contrast, RNA concentrations were measured using a Qubit 4 fluorometer (ThermoFisher Scientific Inc., Waltham, MA, USA). 100 nanograms of total RNA was used as input for ribosomal RNA (rRNA) depletion using the NEBNext rRNA depletion kit (New England BioLabs Inc., Ipswich, MA, USA, product code: E6350), according to the manufacturer’s instructions. Following rRNA depletion, total RNA libraries were built using the NEBNext Ultra II library prep kit from Illumina (New England BioLabs Inc., Ipswich, MA, USA, product code: E7775), according to the manufacturer’s instructions. The yield of amplified libraries was measured on a Qubit 4 fluorometer using a Qubit high-sensitivity DNA kit (HS DNA kit). Amplified libraries were further analyzed on HS D1000 screen tape on a TapeStation system 4150 to assess library size and molarity prior to pooling. The libraries were sequenced using Illumina HISeq-4000.
For circRNA detection, total RNA seq reads were mapped to the mouse GRCm38 genome with BWA [83] (version 0.7.17-r1188) using the -T 19 option. CircRNAs were identified using CIRI2 [84] with default parameters. For further analyses, we used circRNAs identified in at least two replicates at the same time point. To compare circRNAs with circbase datasets, circbase coordinates were translated from mm9 to mm10 genomes using the UCSC liftOver tool. To calculate the p-values and log2 fold changes for circRNA expression changes, we used the DESeq2 [85] (version 1.22.1) package (normalization: rld, test: Wald’s test, p-value adjustment: Benjamini-Hochberg). The circRNA expression table was concatenated with the linear constitutive gene expression table. To calculate the expression of the linear genes, we mapped the reads to the mouse GRCm38 genome with STAR (version 2.7.3.a, default parameters, [86]) and assigned reads to genes with featureCounts [87] (version 2.0.0). For the FPKM calculation, the read counts were normalized to the total number of uniquely mapped reads per sample and the length of the genes, as reported by featureCounts (calculated from the gencode GRCm38 vM12 GTF file). To compare the linear and circular RNA expression changes in the hippocampus and cortex, we quantified the sum of all spliced reads mapped to the gene (featureCounts -J parameter). The evidence plots for Figure 2A,B were produced with R tmod (version 0.46.2 [88]). Gene ontology analysis for Supplementary Table S2 and Figure 2E was carried out with R tmod and msigdbr (7.4.1) packages. | true | true | true |
|
PMC9604106 | 35980185 | Jesús Guzmán-Moreno,Luis Fernando García-Ortega,Lilia Torres-Saucedo,Paulina Rivas-Noriega,Rosa María Ramírez-Santoyo,Lenin Sánchez-Calderón,Iliana Noemi Quiroz-Serrano,Luz Elena Vidales-Rodríguez | Bacillus megaterium HgT21: a Promising Metal Multiresistant Plant Growth-Promoting Bacteria for Soil Biorestoration | 18-08-2022 | Bacillus megaterium,metal resistance,plant growth-promoting bacteria,metal contamination | ABSTRACT The environmental deterioration produced by heavy metals derived from anthropogenic activities has gradually increased. The worldwide dissemination of toxic metals in crop soils represents a threat for sustainability and biosafety in agriculture and requires strategies for the recovery of metal-polluted crop soils. The biorestoration of metal-polluted soils using technologies that combine plants and microorganisms has gained attention in recent decades due to the beneficial and synergistic effects produced by its biotic interactions. In this context, native and heavy metal-resistant plant growth-promoting bacteria (PGPB) play a crucial role in the development of strategies for sustainable biorestoration of metal-contaminated soils. In this study, we present a genomic analysis and characterization of the rhizospheric bacterium Bacillus megaterium HgT21 isolated from metal-polluted soil from Zacatecas, Mexico. The results reveal that this autochthonous bacterium contains an important set of genes related to a variety of operons associated with mercury, arsenic, copper, cobalt, cadmium, zinc and aluminum resistance. Additionally, halotolerance-, beta-lactam resistance-, phosphate solubilization-, and plant growth-promotion-related genes were identified. The analysis of resistance to metal ions revealed resistance to mercury (HgII+), arsenate [AsO4]³–, cobalt (Co2+), zinc (Zn2+), and copper (Cu2+). Moreover, the ability of the HgT21 strain to produce indole acetic acid (a phytohormone) and promote the growth of Arabidopsis thaliana seedlings in vitro was also demonstrated. The genotype and phenotype of Bacillus megaterium HgT21 reveal its potential to be used as a model of both plant growth-promoting and metal multiresistant bacteria. IMPORTANCE Metal-polluted environments are natural sources of a wide variety of PGPB adapted to cope with toxic metal concentrations. In this work, the bacterial strain Bacillus megaterium HgT21 was isolated from metal-contaminated soil and is proposed as a model for the study of metal multiresistance in spore-forming Gram-positive bacteria due to the presence of a variety of metal resistance-associated genes similar to those encountered in the metal multiresistant Gram-negative Cupriavidus metallidurans CH34. The ability of B. megaterium HgT21 to promote the growth of plants also makes it suitable for the study of plant-bacteria interactions in metal-polluted environments, which is key for the development of techniques for the biorestoration of metal-contaminated soils used for agriculture. | Bacillus megaterium HgT21: a Promising Metal Multiresistant Plant Growth-Promoting Bacteria for Soil Biorestoration
The environmental deterioration produced by heavy metals derived from anthropogenic activities has gradually increased. The worldwide dissemination of toxic metals in crop soils represents a threat for sustainability and biosafety in agriculture and requires strategies for the recovery of metal-polluted crop soils. The biorestoration of metal-polluted soils using technologies that combine plants and microorganisms has gained attention in recent decades due to the beneficial and synergistic effects produced by its biotic interactions. In this context, native and heavy metal-resistant plant growth-promoting bacteria (PGPB) play a crucial role in the development of strategies for sustainable biorestoration of metal-contaminated soils. In this study, we present a genomic analysis and characterization of the rhizospheric bacterium Bacillus megaterium HgT21 isolated from metal-polluted soil from Zacatecas, Mexico. The results reveal that this autochthonous bacterium contains an important set of genes related to a variety of operons associated with mercury, arsenic, copper, cobalt, cadmium, zinc and aluminum resistance. Additionally, halotolerance-, beta-lactam resistance-, phosphate solubilization-, and plant growth-promotion-related genes were identified. The analysis of resistance to metal ions revealed resistance to mercury (HgII+), arsenate [AsO4]³–, cobalt (Co2+), zinc (Zn2+), and copper (Cu2+). Moreover, the ability of the HgT21 strain to produce indole acetic acid (a phytohormone) and promote the growth of Arabidopsis thaliana seedlings in vitro was also demonstrated. The genotype and phenotype of Bacillus megaterium HgT21 reveal its potential to be used as a model of both plant growth-promoting and metal multiresistant bacteria. IMPORTANCE Metal-polluted environments are natural sources of a wide variety of PGPB adapted to cope with toxic metal concentrations. In this work, the bacterial strain Bacillus megaterium HgT21 was isolated from metal-contaminated soil and is proposed as a model for the study of metal multiresistance in spore-forming Gram-positive bacteria due to the presence of a variety of metal resistance-associated genes similar to those encountered in the metal multiresistant Gram-negative Cupriavidus metallidurans CH34. The ability of B. megaterium HgT21 to promote the growth of plants also makes it suitable for the study of plant-bacteria interactions in metal-polluted environments, which is key for the development of techniques for the biorestoration of metal-contaminated soils used for agriculture.
Heavy metals are present in the environment due to a variety of natural processes, such as volcanic activity, weathering of rocks, and windblown dust particles (1). However, the production of heavy metal wastes derived from mining and other anthropogenic activities has significantly increased the concentration and bioavailability of toxic metals in ecosystems (1, 2). Some metals (i.e., Zn, Fe, Co, Ni, Cu, Mg, and Mn) are essential at low concentrations for the metabolic activities of living organisms; however, the homeostasis of these metals must be tightly regulated by cells to avoid their toxic effects (3). On the other hand, heavy metals without known biological functions (i.e., Hg, Pb, Ag, Cr, Cd) are highly toxic even at low concentrations for most living organisms, particularly for microbial and plant communities in soils, and constant exposure to toxic metals alters their composition, growth, activity, and genetic variability (4, 5). Despite the toxicity caused by metals, a variety of bacteria that have been constantly exposed to toxic metal ions have tolerance/resistance mechanisms that enable them to survive and colonize hostile environments (6). Bacterial resistance/tolerance to metal ions generally involves extracellular complexation by sorption on the cell surface, intracellular metal accumulation and complexation through intracellular binding to metallothioneins and chelation by siderophores, the intracellular redox of metal ions by enzymatic activity, and cellular exclusion by efflux systems (5, 7). Some metal-resistant bacteria belong to well-described bacterial genera known as plant growth-promoting bacteria (PGPB), which in addition to modifying the toxic effects that metal ions produce in plants (8–10), also promote plant growth in contaminated environments through a variety of mechanisms, including nutrient mobilization, degradation of organic substances, phosphate solubilization, nitrogen fixation, antibiotic production, iron sequestration, and the production of vitamins and phytohormones such as gibberellins and auxins (11–13). The Bacillus genus of bacteria includes a wide variety of species recognized as PGPB (14, 15); among these, Bacillus megaterium (Priestia megaterium) is one of the most studied and is a common inhabitant of soil that is frequently associated with metal-contaminated environments (16, 17). In B. megaterium, the metal tolerance associated with biosorption and siderophore production has been described for metals such as Ni (18), Cd (10, 19, 20), Pb (19, 20), Cu, and Zn (17). However, except for the mercury resistance operon (mer), which has been fully studied (21, 22), few studies of metal resistance determinants and mechanisms in B. megaterium have been reported. These studies include a description of resistance genes for Cd (17, 23, 24), Cu (25), Pb (17), and Ni (26). This work was focused on the characterization of the Hg-tolerant bacterial strain HgT21, isolated from the rhizosphere of a metal-contaminated soil. Genome analysis was carried out to determine its potential for use as a plant growth promoter in metal-contaminated soils on which extreme living conditions are prevalent. Phylogeny and genome analyses and in vitro assays indicate that the isolate HgT21 identified as B. megaterium contains a battery of genes and operons that confer resistance to a variety of heavy metals(oids) and several traits related to halotolerance and plant growth promotion. The results reveal that B. megaterium HgT21 can be used as a model for the study of metal multiresistance in Gram-positive bacteria, with application as a PGPB for phytoimmobilization and restoration of metal-contaminated sites.
The HgII+ tolerance of the HgT21 strain was determined using the MIC for mercuric chloride (HgCl2) in both agar and liquid LB medium. The MIC value in solid medium was 975 μM, whereas in liquid medium, bacterial growth was fully inhibited at 75 μM (Fig. 1).
The whole-genome sequence of strain HgT21 was obtained using the Illumina platform with MiSeq Illumina paired-end technology with 300-bp reads. Our assembly consisted of 26 contigs spanning ~5.5 Mbp with an average GC content of 37.6% and an N50 size of 1,202,273 bp (~1.2 Mb). Genome annotation by Rapid Annotation using Subsystem Technology (RAST) predicted 5,741 protein coding DNA sequences (CDSs) assigned to 351 subsystems. The subsystems feature genes including bacteriocins and antibacterial peptides (n = 9), membrane transport systems (n = 97), auxin biosynthesis (n = 5), nitrogen metabolism (n = 25), metabolism of aromatic compounds (n = 16), iron acquisition and metabolism (n = 65), phosphorus metabolism (n = 45), resistance to antibiotic and toxic compounds (n = 87), and stress response (n = 163). Additionally, the annotation process detected a total of 34 rRNA genes, 114 tRNAs, 117 ribozymes, and 6 prophage regions, 3 of which were identified as active prophages (Fig. 2). No plasmid was identified when analyzed using PLACNETw and Plasmid Finder.
The morphology and biochemical characteristics, such as the bacillus shape, spore formation, positive Gram staining, acid production from glucose, and the use of mannitol and citrate as carbon sources, as well as the presence of catalase, amylase, and oxidase enzymatic activities (Table 1), identified the HgT21 strain as a member of the Bacillus genus based on characteristics reported for this genus (27, 28). Additional enzymatic activities and biochemical characteristics related to carbohydrate, polyalcohol, and xenobiotic metabolism are shown in Table 1.
The HgT21 strain was identified as Bacillus megaterium by whole-genome multilocus sequence typing (wgMLST) analysis. A rooted species tree was inferred for HgT21 and 28 closely related species (see Fig. S1 in the supplemental material) using 8,098 orthogroups. Phylogenetic analysis revealed the close evolutionary relationship of HgT21 with B. megaterium DSM-319 (Fig. S2). In general, the close evolutionary relationship between B. megaterium and Bacillus aryabhattai strains exhibited in the tree is in agreement with previous studies (29, 30). The tree topology groups the strains of B. aryabhattai and B. megaterium into two main clades, with B. megaterium WSH-002 as a common ancestor of these clades.
To determine the genomic plasticity and global gene reservoir of the clade B. megaterium-B. aryabhattai, a pangenome analysis was performed based on the annotated protein sequences of the 29 strains (Fig. S3). The 157,299 protein sequences present across all genomes were clustered into 15,974 orthogroups, representing the pangenome. Among them, 1,633 orthogroups (10.22%) were conserved in all 29 genomes, representing the core genome of all species analyzed (all-core), and 1,229 orthogroups (7.69%) of the core genome retained only one copy in every strain. Moreover, we identified 3,341 orthogroups (20.91%) conserved in the clade B. megaterium-B. aryabhattai (excluding B. aryabhattai PHB10) (MA-core), 3,007 of which (90%) were orthogroups of single-copy genes. On the other hand, 1,954 (12.23%) orthogroups were conserved among the outgroup clade (93.5% single copy) (out-core). To determine the relationship of core genes with some evolutionary features, we classified these genes into different functional categories using Clusters of Orthologous Genes (COG) annotation. We assigned 1,621, 1,309, and 1,166 functions for 80.36%, 66.83%, and 71.4% of proteins in the MA, out, and all-gene cores, respectively. Proteins involved in general function dominated the three cores analyzed (Fig. S4). Some of the most abundant proteins in this class were related to predicted hydrolases of the haloacid dehydrogenase (HAD) superfamily, Zn-dependent hydrolases, including glyoxylases, and predicted permease, a member of the PurR regulon. Through comparison between COG clusters, we found that transcription, secondary metabolite biosynthesis, transport, and catabolism and signal transduction mechanisms were more abundant in the MA core than in the other cores. Translation, ribosomal structure and biogenesis, replication, recombination and repair, and nucleotide transport and metabolism were the least abundant (Fig. S4). The distribution of accessory genes also varied among strains as well as clades (Fig. 3). A total of 7,333 (45.9%) orthogroups were singletons present in only one strain, which was reduced to 490 (3.06%) when considering only the B. megaterium-B. aryabhattai clade. A total of 332 strain-specific genes were identified in B. megaterium HgT21, including 230 annotated hypothetical proteins (Table S2).
B. megaterium HgT21 genome analysis revealed the presence of several genes involved in heavy metal resistance. These genes are associated with resistance to arsenic, copper, mercury, tellurium, zinc, cadmium, and cobalt and are grouped as operons (Fig. 4). Two putative arsenic resistance operons were found (contigs 2 and 8; Fig. 4). The genes arsC, arsB, and arsR and an extra copy of arsR were found at contig 2, whereas arsA, arsD, arsC, arsB, arsR, and arsB-acr3 and an extra copy of arsR were found at contig 8. Additionally, two putative identical copper resistance-like operons (csoR-copZ-copA) were found (contigs 3 and 8; Fig. 4), as well as a putative mercury resistance operon (merRETPA) and the merB gene (contig 8; Fig. 4). A putative tellurium resistance operon that includes telA, yceG, and three tandem copies of telD was found (contig 3; Fig. 4). Finally, three genes associated with Zn-Cd-Co resistance (zntA, arsR, and czcD) were identified. Overall, these genes and operons showed gene arrangements similar to those reported in bacteria, with some differences in the cases of cadmium, arsenic, and tellurium (22, 31–33).
Based on the metal-resistance operons identified in HgT21, resistance to several metal ions (Hg+/Zn2+/Cu2+/Co2+/As3+) was evaluated. The viability of HgT21 bacterial cells in liquid culture exposed to different metal concentrations is shown in Fig. 5. The dose-response curves reveal a gradual decrease in viability when the metal ion concentration increased. Lethal doses of Hg+, Zn2+, Cu2+, Co2+, and As3+ were established at 40 μM, 0.6 mM, 3.0 mM, 420 mM, and 210 mM, respectively.
In accordance with the plant growth promotion-associated genes in HgT21, its ability to promote the growth of Arabidopsis thaliana seedlings was evaluated. The seedlings were exposed through both direct and distant interactions with HgT21 cells for 8 days. After bacterial exposure, a highly branched root system and leafy shoot were observed compared with the control plants (Fig. 6A). Particularly, an increased number and density of lateral roots were observed in both types of interaction (contact and distant), whereas the inhibition of primary root elongation was evident when roots were in contact with bacteria (Fig. 6A and D). Accordingly, both the wet and dry weights of the plants exposed to direct and distant interactions with HgT21 cells were also increased compared to the control (Fig. 6B and C).
Due to the observed ability of the HgT21 strain to induce qualitative and quantitative changes in the growth of A. thaliana plants, the production of IAA (a phytohormone commonly produced by PGPB) was evaluated. IAA synthesis by the HgT21 strain was demonstrated through the Salkowski assay, which revealed that 3 mg/L IAA was produced from tryptophan (Fig. 7).
Based on the results of the genome analysis, the HgT21 strain possesses antibiotic resistance-associated genes. To determine the antibiotic multiresistance of HgT21, which could represent a risk for antibiotic resistance dissemination in natural environments, the antibiotic susceptibility of HgT21 was evaluated. The results in Table 2 reveal resistance to some beta-lactam antibiotics (ampicillin, dicloxacillin, and penicillin and the cephalosporins cefotaxime, ceftazidime, and cefuroxime) and optochin (a hydroquinine derivative). Susceptibility to beta-lactam cephalosporins (cephalothin, cefepime, and ceftriaxone) and to different groups of antibiotics, including vancomycin (glycopeptide), nitrofurantoin (sulfamide), erythromycin (macrolide), levofloxacin, pefloxacin and ciprofloxacin (quinolones), gentamicin, amikacin and netilmicin (aminoglycosides), polymyxin and bacitracin (polypeptides), novobiocin (aminocumarine), tetracycline, chloramphenicol, and trimethoprim/sulfamethoxazole, was also observed.
In recent decades, sustainable agriculture for safe food production has become an important challenge for soil chemistry, plant biology, agricultural and environmental engineering, and soil microbiology due to the increased demand for healthy safe products (11). Heavy metal pollution in agricultural soils leads to the accumulation of toxic metals in plants and, consequently, their integration into the food chain. Thus, the toxicity of heavy metals for most living organisms at relatively lower concentrations (at the level of ppb or ppm) represents a threat to human health, safe food production, and maintenance of microbial diversity in soils (1, 34). The development of strategies for the recovery of degraded and metal-contaminated agricultural soils could be based on bioremediation approaches using beneficial microorganisms such as metal-resistant PGPB and natural pathogen antagonists as an alternative for metal immobilization or as a replacement for chemical fertilizers and pesticides to guarantee sustainable agriculture (12, 14, 15). In this context, the search for soil plant-associated PGPB and metal biotransformation bacteria is important for the establishment of sustainable strategies for agricultural soil restoration. In this work, a metal-resistant PGPB was isolated and characterized. Based on 16S and complete-genome sequence analyses, the isolated HgT21 was identified as Bacillus megaterium. The phylogenetic tree based on the genome analysis shows that B. megaterium HgT21 is closely related to B. megaterium DSM319 and is grouped within the B. aryabhattai-B. megaterium clade. In addition, the average nucleotide identity (ANI) values (>95 to 96%) between the B. aryabhattai and B. megaterium strains (Fig. S1) support the proposition of Narsing Rao et al. to reclassify B. aryabhattai as the later heterotypic synonym of B. megaterium de Bary 1884 (35) due to the ANI values being higher than the recognized threshold values for bacterial species delineation (36). As in B. megaterium DSM319, no plasmid sequences were identified; however, a low GC content was observed in some contigs (6 and 8 to 11), and the presence of active prophages (contigs 10 and 11; Fig. 2) and higher coverage (>80×; contigs 6, 8, 10, and 11) suggest the integration of plasmid-borne genes into genomic DNA and/or the exchange of genes between plasmids and chromosomes. This notion is supported by studies that provide evidence of extensive gene transfer between the plasmids of B. megaterium QMB1551 and its own chromosome and those of the plasmid-less B. megaterium DSM319 (37). According to genotype identification, the morphology of the HgT21 cells correlates with the B. megaterium phenotype initially described by De Bary in 1884 (Gram-positive, approximately 8 to 10 μm size, and spore-forming), and the identification was also corroborated based on biochemical characteristics reported for B. megaterium (27). Bacillus megaterium is widely recognized as a soil bacterium with potential industrial applications due to its ability to utilize different carbon sources, grow in a wide range of temperatures (3 to 45°C), produce proteases, promote plant growth, and antagonize plant pathogens for biocontrol (38–40). In agreement with these reports, genome annotation based on protein prediction revealed that B. megaterium HgT21 contains approximately 264 genes related to the production of bacteriocin and antibacterial peptides, resistance to antibiotic and toxic compounds, auxin biosynthesis, and stress response. Moreover, the biochemical profile of B. megaterium HgT21 reveals its ability to use diverse carbon sources (Table 1). These results are consistent with key enzyme genes encountered in its genome, which are involved in the metabolism of these carbohydrates and polyalcohols (Table 2). Additional enzymatic activities detected in B. megaterium HgT21 are also correlated with the presence of key genes involved in these enzymatic activities (Table 2), except for urease; in contrast to the biochemical data, urease was identified in the B. megaterium HgT21 genome, suggesting dysfunction of this enzyme. According to the recognition of the Bacillus genus as PGPB, genes involved in one of the three metabolic pathways described for the synthesis of indole-3-acetic acid (IAA), a nonvolatile phytohormone involved in plant growth promotion (41), were found in B. megaterium HgT21. The production of IAA by PGPB is related to the synthesis pathway of tryptophan. In B. megaterium HgT21, the genes required for IAA synthesis through the chorismate pathway were identified (Table 2) and involve the conversion of chorismate to tryptophan (by enzymes encoded in the trpABCDEG operon) and its subsequent conversion to IAA by nitrilase activity. These results suggest that enzymes encoded by genes of the chorismate pathway could be involved in IAA production by B. megaterium HgT21. Moreover, genes involved in the synthesis of volatile organic compounds (VOCs) such as acetoin and butanediol were also identified; these genes encode enzymes that convert pyruvate to acetoin (acetolactate synthase and acetolactate decarboxylase) and acetoin to 1,3-butanediol (2,3-butanediol dehydrogenase [BDH] and glycerol dehydrogenase [GDH]) (42). Although the activity of these enzymes was not demonstrated in this work, the presence of their encoding genes in the HgT21 genome could be related to the increased growth of A. thaliana seedlings observed after its exposure to both distant and contact interactions with B. megaterium HgT21. The phenotype observed in A. thaliana after the interaction with B. megaterium HgT21 correlates with previous reports that describe the increase of lateral roots, decreasing cell elongation in primary root, and root hair density increase upon colonization (43). In relation to the phenotype observed, it has been described that some PGPB indirectly reduce the effect of ethylene production (an inhibitor of primary growth root) through the production of the ACC (1-aminocyclopropane-1-carboxylate) deaminase, which degrades ethylene, and in consequence induces primary root elongation (44). However, the gene encoding the ACC deaminase was not found in B. megaterium HgT21; this could explain the inhibition of the root growth when colonization is established. According to reports that demonstrate that the production of VOCs by rhizobacteria induces the elongation of lateral roots and root hairs of A. thaliana seedlings (45), the production of acetoin and 2,3-butanediol in HgT21 could be responsible for the observed elongation of lateral roots and root hairs of seedlings; however, this must be experimentally demonstrated, and future experiments must be conducted to evaluate the contribution of VOCs and diffusible compounds to growth promotion. Additionally, the group of genes involved in phosphate solubilization (41, 46) that were identified in B. megaterium HgT21 reveal its potential as a phosphate solubilizer bacterium. Bacillus megaterium is also recognized for its adaptation and resistance to stressful environmental conditions such as saline or acidic soils and soil contamination with heavy metals and other xenobiotics (40, 47, 48). Accordingly, B. megaterium HgT21 showed high tolerance to HgII+ ions, and the MIC values were higher in solid (975 μM HgCl2) than in liquid (75 μM HgCl2) medium, presumably due to the major interaction of the bacterial cell surface with metal ions in liquid medium (49, 50). This result correlates with observations in mercury-resistant Bacillus species isolated from Minamata Bay, Japan, a highly mercury-contaminated site, which have MIC values from 80 to 320 μM HgCl2, particularly the well-characterized mercury-resistant strain B. megaterium MB1 (MIC of 80 μM HgCl2). These values are significantly higher than those of the mercury-sensitive strain (no mer determinants) B. megaterium WH20 (MIC of 10 μM HgCl2) (16, 22). Genomic analysis revealed that B. megaterium HgT21 contains a variety of heavy metal resistance-associated genes and/or complete putative operons, including the merRETPA operon and the merB3 gene, which confer a wide spectrum of mercury resistance (51). The mercury resistance system in HgT21 showed a gene arrangement identical to that of the mer-like determinant in Tn5083 of B. megaterium MK64-1 (Kamchatka, Russia), which has been proposed to be a derivative of mer determinants reported in Tn5084 of Bacillus cereus RC607 (Boston, USA) and TnMERI1 of B. megaterium MB1 (Minamata, Japan) (21, 22, 52–54). These results support the idea that recombination events of transposition gene exchange could be an important contribution to the evolution of HgII+-resistant transposons in Gram-positive bacteria and the worldwide horizontal dissemination of the class II TnMERI1-like transposons across bacterial species and geographical barriers (22, 55). Bacillus megaterium HgT21 also contains a variety of metal(oid) resistance systems, which include well-described genes and operons for arsenic, copper, cadmium, zinc, cobalt, and tellurium resistance. Two arsenic resistance-like operons were identified in HgT21. First, the canonical operon arsRBC (contig 8), initially described in Staphylococcus aureus (pI258 plasmid), correlates with its wide distribution in the plasmids and chromosomes of Bacteria and Archaea from different origins (32). However, in Bacillus species, only variants of this operon have been reported, as well as the arsR2-orf2-arsB-arsC2 operon in the chromosome of Bacillus sp. strain CDB3 and the arsR-orf2-acr3-arsC operon in the “Tn (skin element)” of Bacillus subtilis JH642 (31, 56). The second arsenic resistance-like operon (arsRBCDA) identified (contig 2) showed an arrangement similar to that of the commonly reported operon arsRDABC (57) and was identical to those reported in the chromosome of Bacillus sp. CDB3 (31). The presence of two arsenic resistance operons, extra copies of arsR (upstream from arsRBC and arsRBCDA) and acr3 (upstream from arsRBCDA), correlates with the wide complexity and variety of gene configurations in ars clusters reported in prokaryotes (32, 58). The enzymes encoded by the arsenate resistance-like operons in HgT21 could be responsible for the observed arsenate resistance in this strain, which is similar to the most arsenic-resistant Bacillus species (56, 59, 60). The presence of the acr3 gene (arsB homolog) in Bacillus megaterium HgT21 contrasts with studies about the prevalence of acr3 in different orders of Bacteria, which revealed that acr3 is mainly present in Burkholderiales, Actinobacteria, and Alphaproteobacteria, whereas arsB is prevalent in Firmicutes and Gammaproteobacteria (61). Taken together, the results support studies that revealed that the presence of multiple and redundant ars genes in bacteria exposed to selective pressure is the result of gene duplication via horizontal transfer (32). The arrangement of the three copper resistance-associated genes in B. megaterium HgT21 (csoR-copZ-copA) is identical to the copper resistance system described in B. subtilis (33), which includes the transcriptional regulator CsoR (a widespread copper-inducible repressor distributed in Gram-positive bacteria and Proteobacteria that regulates expression of the copZA operon), the copper chaperone CopZ, and the copper-efflux ATPase CopA (62). The presence of this copper-like resistance system in B. megaterium HgT21 correlates with the identification of this system in approximately half of the members of Firmicutes, including the Bacillales “copper users,” for which copper is essential for aerobic respiration, acting as a cofactor in terminal enzymes of the aerobic pathway (63). In copper users, the csoR-copZA system confers resistance to high levels of copper (33), as was observed in B. megaterium HgT21 (3 mM CuSO4), compared with the copper resistance reported in both Gram-positive and Gram-negative bacteria (3.5 to 5.5 mM CuSO4) (12, 33, 64). The presence of two identical copper resistance systems at different locations in the genome suggests duplication of this system associated with mobile elements. Three genes associated with resistance to different metal cations (Zn2+, Cd2+, Pb2+, and Co2+) were identified in B. megaterium HgT21. cadA encodes a cadmium-translocating P-type ATPase (CadA), a homologous protein of ZntA, which is considered a multipurpose metal-exporting pump for the extrusion of Zn2+, Cd2+, Ag2+, and Pb2+ (65). The cadA gene is a component of the well-described cadmium resistance operon cadAC encountered in Gram-positive bacteria, including S. aureus (plasmid pI258), Listeria monocytogenes, B. megaterium, and other Bacillus species (24, 66–68). In the CadAC system, the expression of cadA is tightly controlled by the regulatory protein CadC, a Cd2+/Pb2+/Zn2+ responsive repressor encoded by cadC located downstream of cadA (67). Interestingly, although cadA was located in B. megaterium HgT21, cadC was not found; instead, a hypothetical protein was located upstream of cadA. However, in the opposite direction, near and upstream of cadA, the genes arsR (encoding an ArsR family transcriptional regulator) and czcD (encoding a membrane-bound protein member of the metal-diffusion facilitator, cation diffusion facilitator (CDF) subfamily) were found. This suggests that in the absence of CadC (a homodimeric repressor that belongs to the ArsR/SmtB family of metalloregulatory proteins), ArsR could act as a transcriptional regulator of cadA (67). With respect to czcD, it has been reported that CzcD is a heavy metal transporter involved in the regulation of the czc (cadmium, zinc, cobalt) resistance system described in Ralstonia sp. strain CH34 (Cupriavidus metallidurans CH34) (69, 70); however, the absence of additional czc genes in B. megaterium HgT21 suggests that, rather than acting as a regulator protein, CzcD could mediate resistance against Zn2+/Co2+/Ni2+/Cd2+ through an antiporter mechanism catalyzing the active efflux of divalent ions in exchange for K+/H+, as described in Bacillus subtilis (71, 72). Thus, czcD could play a physiological role related to the maintenance of metal homeostasis, as has been proposed for members of CDF family, such as CzcD in Streptococcus pneumoniae, ZRC-1 in S. cerevisiae and Znt-1 and Znt-2 in mammals (73–75). In agreement with this result, the presence of czcD without other components of the czc and cadAC operon was described in Bacillus megaterium (26) and Bacillus subtilis (71). In Bacillus paranthracis, both czcD and cadA are present; however, the gene arrangement and genome location have not been described (23). Thus, to our knowledge, this is the first report of a genic arrangement that includes cadA, arsR, and czcD, each corresponding to three different metal-resistance systems (czc, cadAC, ars). CadA and CzcD are involved in the extrusion of divalent metal ions, suggesting that this chimeric system probably arises for the resistance to and/or homeostasis maintenance of divalent metal cations (Zn2+, Cd2+, Pb2+, and Co2+) in B. megaterium HgT21, which showed Zn2+ resistance (0.6 mM) comparable to that reported for zinc-resistant bacteria (0.1 to 1.0 mM) (65, 71); however, this must be experimentally demonstrated. Interestingly, the Co2+ resistance was extremely high (420 mM) relative to the Co2+ resistance reported in bacteria (0.425 to 8 mM CoCl2) (70, 76), and the mechanism involved in this unprecedented level of Co2+ resistance is under investigation. A set of genes associated with tellurium resistance (TeR) was also identified in B. megaterium HgT21. TeR determinants are a group of ubiquitous genes distributed across phylogenetically diverse taxa in bacteria isolated from diverse environments (77). Based on the high sequence similarity, it has been proposed that TeR genes arose from a common ancestor that evolved as the result of adaptation to ancient metal-rich environments or through horizontal transfer events (77, 78). A variety of TeR genes have been identified in plasmids and chromosomal genomic islands (GIs) of pathogenic and extreme-environment isolated bacteria; however, the function of proteins encoded by these genes remains unclear (77, 79). The operon terZABCDE is commonly reported in bacteria; however, the minimal fragment required to confer tellurite resistance is terBCDE, whereas the complete operon also provides phage resistance and pore-forming colicins, explaining its maintenance in a variety of pathogenic bacteria for which tellurite resistance is probably not necessary (80). Moreover, comparative genomics revealed that ter genes are functionally linked to enzymes involved in DNA processing and repair and induced by H2O2 and superoxide, suggesting their association with the general stress response caused by ROS production or tellurite ions (81, 82). In this context, telA, yceG, and the three tandem copies of terD in B. megaterium HgT21 could be associated with different functions: (i) tellurite resistance, considering that it was isolated from a metal-polluted environment and considering the role of TelA in tellurite resistance in Rhodobacter sphaeroides (83), and/or (ii) to deal with hostile environmental conditions related to ROS generation, considering that yceG and terD have been associated with the cellular defense against oxidative stress in pathogenic bacteria (82, 84). However, the function of enzymes encoded in tellurite resistance-related genes in B. megaterium HgT21 must be experimentally demonstrated. Due to bacterial metal resistance being associated with antibiotic resistance (85, 86), the antibiotic susceptibility in B. megaterium HgT21 was assessed. Susceptibility to most of the antibiotic families was observed, suggesting a low risk of antibiotic resistance transfer in the environment. Resistance to some beta-lactams correlates with the presence of genes associated with beta-lactam resistance. Optochin resistance could probably be produced by a lack of recognition of the target site of this antibiotic (membrane ATPase F0F1), which can be altered by mutations in the ATPase F0F1 gene (87). The resistance to beta-lactams, bacteriocins, and antibacterial peptides in B. megaterium HgT21 could be advantageous in its natural hostile environment, in which competence with other microorganisms could be crucial to survive. The Bacillus species in the soil habitat are well recognized as “zymogenous” bacteria and ecologically have been defined as “r-strategists,” which means that they can grow quickly when the nutrient supply is abundant, as in the rhizosphere, and possess a high colonization and competitive ability. Importantly, quorum sensing-mediated processes in Bacillus species, such as endospore and biofilm formation, constitute an important survival strategy in soil under nutrient-limited conditions and hostile environments; its high adaptability to the environment is evidenced by its ubiquity in both nonextreme and extreme soils (88). In this context, the genetic information of this bacterium and the preliminary physiological trait analysis (metal multiresistance, plant growth promotion, IAA production, use of carbon sources) suggest a high adaptability of B. megaterium HgT21 to adverse environments such as metal-contaminated soils and its potential as a PGPB through phytohormone and VOC production, phosphate solubilization, and antibiotic production. Considering the essential characteristics that help to define new PGPB as biofertilizers (high rhizosphere competence, ability to increase plant biomass, long-term survival, plant-beneficial physiological traits, a lack of risk factors for human and environmental health, and high tolerance to environmental stresses encountered in soil/plant systems) (88), the genetic information of the HgT21 strain suggests that it could be suitable for use as a biofertilizer in the promotion of plant growth in metal-contaminated environments due to its ability to produce/perform quorum sensing-related processes, endospores, flagella, siderophores, metabolism of a variety of organic compounds, IAA production, and resistance to a variety of metals. However, in vivo experiments under more realistic conditions must be conducted to demonstrate that it can be used as a good biofertilizer. Moreover, due to its sporulating, nonpathogenic, and free-endotoxin nature, B. megaterium HgT21 could be a good candidate for biotechnology applications in the food and pharmaceutical industries, among others (38).
The mercury-tolerant bacterial strain HgT21 was isolated from the rhizosphere of Fabaceae plants (Dalea bicolor) inhabiting mining tails located in Zacatecas, Mexico (22°47′01.4″ N, 102°36′21″ W and 2,426 meters above sea level [m.a.s.l.]). To test the HgT21 tolerance to mercury ions (HgII+) in solid medium, the MIC was determined. Briefly, LB agar was supplemented with HgCl2 at final concentrations of 0, 1, 10, 100, 200, 400, 800, 850, 900, 925, 950, 975, and 1,000 μM and poured into petri dishes. After agar solidification, the plates were inoculated with 1 × 105 colony forming units (CFU) and spread onto the agar surface and incubated at 37°C until visible growth of bacteria was detected. The concentration at which no growth was detected was reported as the MIC in solid medium. Mercury tolerance in liquid medium was also determined, by assessing the MIC value as follows: overnight cultures were diluted 1:100 in fresh LB medium supplemented with HgCl2 at final concentrations of 0, 1, 10, 25, 50, 75, and 100 μM and grown at 37°C and 200 rpm for 8 h (when the end of exponential growth was reached in cultures not exposed to mercury ions), and then the optical density was spectroscopically determined at 600 nm (OD600). Three independent experiments were carried out.
This analysis was based on colony characteristics, Gram staining, motility, and biochemical properties such as H2S production, acid and acetylmethylcarbinol production from glucose, use of citrate and mannitol as a carbon source, and the presence of enzymatic activities such as urease, lysine and ornithine decarboxylases, tryptophanase, cytochrome oxidase, catalase, and amylase (89). Based on this characterization, the HgT21 strain was identified according to Bergey’s Manual of Determinative Bacteriology (27). An additional test for the biochemical analysis was performed using an automated Vitek 2 system (bioMérieux) with the GP-ID card for Gram-positive bacteria.
Genome sequencing of Bacillus megaterium HgT21 was performed at Langebio genomics services (Irapuato, Mexico City) using the 2 × 300-bp Illumina MiSeq platform. Raw reads were quality assessed using cutadapt v1.9.1 (90) and Trimmomatic v0.36 (91) to remove adapter sequences and low-quality reads, respectively. De novo assembly of the filtered reads was performed using SPAdes v3.9.1 (92) with the –careful parameter for read error correction. The assembled contigs were annotated using the Rapid Annotation using Subsystem Technology (RAST) web server (http://rast.nmpdr.org) (93) using the ClassicRAST annotation scheme, FIGfams v90, automatic error correction, and automatic frameshift correction. Finally, noncoding RNAs (ncRNAs), tRNAs, and rRNAs were searched in the HgT21 assembly using Infernal v1.1.3 (94) against the Rfam database v14.1 (95). Prophage regions were bioinformatically predicted with Prophage Hunter (96). Only regions with scores of >0.8 were classified as active prophage regions. The presence of plasmid sequences was verified using PLACNETw (97) and PlasmidFinder v2.1 (98). Colinear blocks between contigs of the B. megaterium HgT21 genome were identified using the SibeliaZ-LCB (99) algorithm with default parameters. Gene and ncRNA densities, colinearity blocks, GC content, and prophage regions were visualized in Circos (100). The gene regions were visualized using the R package gggenes v0.4.1 (https://github.com/wilkox/gggenes).
We used a whole-genome multilocus sequence typing (wgMLST) method (101) to identify the HgT21 strain. Protein sequences of 28 species phylogenetically closely related to HgT21 were used for phylogenetic analysis (Table S1). The selection of representative species was carried out according to their pairwise average nucleotide identity (ANI) values, which were determined by FastANI (102). The species tree was inferred using the STAG algorithm and rooted with the STRIDE algorithm in the OrthoFinder program (103). Bacillus cereus B4264 was used as an outgroup.
Gene orthogroups were derived from OrthoFinder analysis. The gene count of all orthogroups was converted to a 0/1 matrix (1 indicates the presence of the gene in the orthogroup, and 0 indicates its absence). The resulting binary matrix was used to generate the gene accumulation curves of the pan- and core genomes using PanGP software (104) with the distance guide method, sample size of 5,000, 100 sample repeats, and 100 amplification coefficients. The species tree and panmatrix were plotted together using the phytools package (105). In addition, the proteins of the pangenome were grouped into one of the three studied cores, B. megaterium-B. aryabhattai, outgroup, and all species, after they were mapped to different COGs using the WebMGA server (106) and RPS-BLAST program.
The survival of HgT21 cells exposed to different metal salts was evaluated as a measurement of their resistance to different metal ions. Lethal doses (LDs; concentration at which the bacterial cells were dead) of metal ions were established throughout viability assays. Bacterial cells were exposed to different metal salts, including sodium arsenite, sodium arsenate, mercury chloride, zinc sulfate, copper sulfate, and cobalt chloride. Briefly, 1 mL of HgT21 cells at the mid-log phase of growth in LB medium (1 × 108 CFU/mL) was deposited in 24-well polystyrene plates (Nunclon Delta surface, Nunc) and treated with metallic salts at final concentrations of 0, 2.5, 5.0, 7.5, 10, 20, 30, 40, and 50 μM HgCl2; 0.1, 0.2, 0.3, 0.4, 0.5, and 0.6 mM ZnSO4; 0, 0.5, 1.0, 1.5, 2.0, and 3 mM CuSO4; 0, 60, 120, 180, 240, 300, 360, and 420 mM CoCl2; or 0, 5, 10, 30, 60, 90, 120, 150, 180, and 210 mM Na3AsO4. The plate was incubated for 2 h at 37°C and 120 rpm. Subsequently, 0.1 mL of each well was taken for serial dilution, and 0.1 mL of each dilution was plated in triplicate onto LB agar and incubated at 37°C for 18 h. The colonies formed were counted and graphically reported as CFU per milliliter (CFU/mL) as a measure of the bacterial viability as a function of the metal ion concentration. Finally, the LD of each metal ion was determined.
Seeds of Arabidopsis thaliana ecotype Col-0 were surface-sterilized with 95% (vol/vol) ethanol solution for 5 min and bleach (20% [vol/vol]; equivalent to 0.13 M sodium hypochlorite) for 7 min. Subsequently, the seeds were washed five times with sterile distilled water and stored at 4°C for 48 h in the dark to promote and synchronize germination. The stratified A. thaliana seeds were placed and germinated on petri dishes containing 0.2× Murashige and Skoog (MS) medium, pH 5.7, and 1% agar TC (micropropagation grade, PhytoTechnology Laboratories) in a horizontal arrangement on the upper zone (eight plants in each plate with 2 cm interdistance) before bacterial inoculation (107). Plates were sealed with plastic film to avoid seed contamination and placed vertically with an angle of 65° in a growth chamber (Percival Scientific) at 22 to 20°C with a photoperiod of 16 h of light and 8 h darkness. The plant growth promotion of the HgT21 strain on A. thaliana was carried out by in vitro direct (in contact) and indirect (distant) interactions of plant root tips with bacterial cells as previously described (108). For this analysis, 6-day-old plates with A. thaliana seedlings were inoculated by streaking with 10 μL (approximately 1 × 106 CFU/mL) of a 24-h bacterial culture of the HgT21 strain using an inoculating loop to draw a straight line over the middle (contact) or the lower zone (distant) of plates containing the A. thaliana seedlings. The inoculation of the lower zone avoids the interaction of plant root tips with bacterial cells (distant), whereas the inoculation in the middle zone allows for the interaction between bacterial cells and plant roots (contact). Petri dishes without bacterial inoculation were also prepared and used as controls. After bacterial inoculation, the plates were incubated in a growing chamber at 22 to 20°C with a photoperiod of 16 h of light and 8 h of darkness. Every 2 days, until day 8 postinoculation, photographs of plates were taken to determine primary root length, number and density of lateral roots (density = number of lateral roots/length of primary root) using ImageJ software (http://rsb.info.nih.gov/nih-image/). In addition, seedlings were removed from the plates, weighed, and dried at 80°C in paper bags until they reached a constant weight. The wet and the dry biomass was determined. All experiments were replicated at least three times. The differences in the number and density of lateral roots, primary root length, and the wet and dry weights of plants exposed and not exposed (control) to direct and distant interactions with HgT21 were calculated by performing one-way analysis of variance (ANOVA) in Microsoft Excel with significance set at P < 0.05.
A culture of HgT21 at the mid-log phase of growth was prepared in LB medium, and 0.1 mL of this culture (containing 1 × 108 CFU/mL) was inoculated in 5 mL of M9 minimal medium supplemented with 0.5% glucose and 0.1% l-tryptophan and incubated for 96 h at 37°C and 180 rpm in triplicate. The amount of IAA produced was determined by the colorimetric method described by Gordon and Weber (109).
Resistance to different antibiotics was tested by the disk diffusion method and interpreted according to the manufacturer’s instructions for BBL antibiotic SensiDiscs (Becton, Dickinson Microbiology Systems, Cockeysville, MD) or Multidiscos (Bio-Rad, Mexico).
The raw sequence reads have been deposited in the NCBI Sequence Read Archive (SRA; SRR17752569 and SRR17752570) under BioProject accession number PRJNA800475. The draft genome sequence was submitted to GenBank with the accession number JAKKUX000000000. | true | true | true |
PMC9604158 | 36169448 | Concha Ortiz-Cartagena,Laura Fernández-García,Lucia Blasco,Olga Pacios,Inés Bleriot,María López,Rafael Cantón,María Tomás | Reverse Transcription-Loop-Mediated Isothermal Amplification-CRISPR-Cas13a Technology as a Promising Diagnostic Tool for SARS-CoV-2 | 28-09-2022 | COVID-19,SARS-CoV-2,RT-LAMP,CRISPR-Cas13,CRISPR-Cas,diagnosis | ABSTRACT At the end of 2019, a new coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), caused a pandemic that persists to date and has resulted in more than 6.2 million deaths. In the last couple of years, researchers have made great efforts to develop a diagnostic technique that maintains high levels of sensitivity and specificity, since an accurate and early diagnosis is required to minimize the prevalence of SARS-CoV-2 infection. In this context, CRISPR-Cas systems are proposed as promising tools for development as diagnostic techniques due to their high specificity, highlighting that Cas13 endonuclease discriminates single nucleotide changes and displays collateral activity against single-stranded RNA molecules. With the aim of improving the sensitivity of diagnosis, this technology is usually combined with isothermal preamplification reactions (SHERLOCK, DETECTR). Based on this, we developed a reverse transcription-loop-mediated isothermal amplification (RT-LAMP)-CRISPR-Cas13a method for SARS-CoV-2 virus detection in nasopharyngeal samples without using RNA extraction that exhibits 100% specificity and 83% sensitivity, as well as a positive predictive value (PPV) of 100% and negative predictive values (NPVs) of 100%, 81%, 79.1%, and 66.7% for cycle threshold (CT) values of <20, 20 to 30, >30 and overall, respectively. IMPORTANCE The coronavirus disease 2019 (COVID-19) crisis has driven the development of innovative molecular diagnosis methods, including CRISPR-Cas technology. In this work, we performed a protocol, working with RNA extraction kit-free samples and using RT-LAMP-CRISPR-Cas13a technology; our results place this method at the forefront of rapid and specific diagnostic methods for COVID-19 due to the high specificity (100%), sensitivity (83%), PPVs (100%), and NPVs (81% for high viral loads) obtained with clinical samples. | Reverse Transcription-Loop-Mediated Isothermal Amplification-CRISPR-Cas13a Technology as a Promising Diagnostic Tool for SARS-CoV-2
At the end of 2019, a new coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), caused a pandemic that persists to date and has resulted in more than 6.2 million deaths. In the last couple of years, researchers have made great efforts to develop a diagnostic technique that maintains high levels of sensitivity and specificity, since an accurate and early diagnosis is required to minimize the prevalence of SARS-CoV-2 infection. In this context, CRISPR-Cas systems are proposed as promising tools for development as diagnostic techniques due to their high specificity, highlighting that Cas13 endonuclease discriminates single nucleotide changes and displays collateral activity against single-stranded RNA molecules. With the aim of improving the sensitivity of diagnosis, this technology is usually combined with isothermal preamplification reactions (SHERLOCK, DETECTR). Based on this, we developed a reverse transcription-loop-mediated isothermal amplification (RT-LAMP)-CRISPR-Cas13a method for SARS-CoV-2 virus detection in nasopharyngeal samples without using RNA extraction that exhibits 100% specificity and 83% sensitivity, as well as a positive predictive value (PPV) of 100% and negative predictive values (NPVs) of 100%, 81%, 79.1%, and 66.7% for cycle threshold (CT) values of <20, 20 to 30, >30 and overall, respectively. IMPORTANCE The coronavirus disease 2019 (COVID-19) crisis has driven the development of innovative molecular diagnosis methods, including CRISPR-Cas technology. In this work, we performed a protocol, working with RNA extraction kit-free samples and using RT-LAMP-CRISPR-Cas13a technology; our results place this method at the forefront of rapid and specific diagnostic methods for COVID-19 due to the high specificity (100%), sensitivity (83%), PPVs (100%), and NPVs (81% for high viral loads) obtained with clinical samples.
Since their emergence at the beginning of the 21st century, coronaviruses have been recognized as a health concern because of their ability to cause severe respiratory infections in humans. At the end of 2019, a new coronavirus appeared, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), producing a novel illness, coronavirus disease 2019 (COVID-19), and showing two remarkable characteristics: the virus causes the development of an unusual viral pneumonia, and it is highly transmissible and thus spreads rapidly (1–3). This led to the SARS-CoV-2 pandemic, which persists to date and has caused more than 6.2 million deaths (WHO COVID-19 Dashboard [https://covid19.who.int/]). Fortunately, vaccination campaigns have decreased the incidence of COVID-19 (4). However, specialists claim that this virus is likely to coexist with us for a long time, as the price of vaccines and the necessary cold-chain stability make it difficult for the vaccine to reach the most remote places in the world, as SARS-CoV-2 does. Together with the fact that no efficient therapy has been developed for COVID-19, this indicates that accurate and early diagnosis in point-of-care (POC) testing is required to minimize the prevalence of SARS-CoV-2 infection (1–3). In the last couple of years, researchers have made great efforts to develop a diagnostic technique that maintains high levels of sensitivity and specificity, without the need for expensive equipment or highly trained personnel for its implementation. Such a diagnostic technique would allow the detection of SARS-CoV-2 infection in health centers, as well as at home or in the field, which would accelerate the identification of infected patients, enabling prompt treatment and halting the spread of SARS-CoV-2 worldwide (5). The use of nucleic acids as biomarkers has become the diagnostic gold standard, because of the species specificity of the technique and because DNA and RNA can be amplified (6). Although the reverse transcription-PCR (RT-PCR) assay is routinely used as the gold standard diagnostic test for COVID-19 (5, 7–10), throughout the pandemic period, it has shown sensitivity levels of 45% to 60% (10) and even lower than 40%, according to some authors (7), and worrying false-negative rates of 2% to 29% (10, 11). Additional downsides of this amplification method are the elevated costs (expensive equipment for implementation and readout of results), the need for specialized personnel in laboratories and the time required (4 to 6 h) (5, 8, 10). Consequently, isothermal amplification reactions are becoming especially important in the diagnosis of COVID-19 (5). Although different methods of isothermal amplification are available, recombinase polymerase amplification (RPA) and loop-mediated isothermal amplification (LAMP) reactions are the methods most commonly used in research. The LAMP-based technique has displayed greater specificity than RPA (5, 12). LAMP has previously been used to detect several microorganisms, and the aforementioned advantages led to its optimization for COVID-19 diagnosis, and it has been applied in association with other techniques which increase diagnostic specificity, such as clustered regularly interspaced short palindromic repeat (CRISPR)-associated protein (CRISPR-Cas) systems (5, 13–15). Naturally, CRIPSR-Cas systems provide adaptive immunity for bacteria and archaea, as they collect genomic fragments (spacers) from foreign elements (bacteriophages, plasmids, and other mobile genetic elements) that are expressed in an RNA molecule form (crRNA) that guides an endonuclease protein (Cas) to the pathogen for the final degradation of its nucleic acid material (16, 17). Since their discovery, CRISPR-Cas systems have revolutionized the field of molecular biology. Initially, they were presented as highly specific tools for genome editing. However, they are also applicable for the diagnosis and treatment of infectious diseases and are now considered key for development in these areas (16, 17). Class 2 CRISPR-Cas systems have a simpler effector structure, which makes them more attractive for use in genome editing, diagnosis, and treatment. In this class, Cas12 and Cas13 proteins display nonspecific endonuclease activity when activated (collateral activity) against single-stranded DNA (ssDNA) and RNA (ssRNA), respectively. This feature could be applied in clinical diagnosis, taking advantage of the reporter molecule target of this activity (collateral-based detection), which acts by amplifying the detection signal. Therefore, Cas12 and Cas13 are proposed as the most promising tools for use in diagnostic techniques, with the latter being particularly important in terms of specificity, as it has the ability to discriminate single nucleotide changes (16). Researchers recently developed two novel assays for detecting SARS-CoV-2 based on CRISPR-Cas technology: DETECTR and SHERLOCK. The DETECTR technique uses reverse transcription-LAMP (RT-LAMP) for amplification and Cas12 as an endonuclease, while SHERLOCK uses RT-RPA for amplification and Cas13 (18, 19). On the basis of these works, in this study, we describe the development and optimization of a LAMP-CRISPR-Cas13a technique for the diagnosis of SARS-CoV-2 infection in clinical samples in a process that does not require RNA extraction or purification (Fig. 1). With this technique, high levels of sensitivity and specificity, comparable to those associated with RT-PCR, were obtained.
We obtained an output of more than 7,000 articles as a result of a search using the keywords “RT-PCR diagnosis COVID-19”, of which almost 4,000 were published in 2021 alone. This result was compared with the findings of Bhatt et al. (20) concerning papers related to RT-LAMP and CRISPR for SARS-CoV-2 diagnosis. Of these, we analyzed 10 articles on the RT-LAMP technique and 10 articles related to RT-LAMP-CRISPR-Cas technology, finding that only 1 applied the endonuclease Cas13 for SARS-CoV-2 diagnosis, but always on samples treated with an RNA extraction kit (21) (Table 1). Data collected from the RT-LAMP articles were used to determine the range of values of the parameters considered: sensitivity, 81% to 98%; specificity, 36% to 100%; positive predictive value (PPV), 86% to 100%; and negative predictive value (NPV), 78% to 99% (Table 1). The results showed that major efforts have been made to detect SARS-CoV-2 in RNA-purified samples (8/10), although RNA extraction-free research has also yielded potentially useful results (sensitivity, >94%; specificity and PPV, 100%; NPV, >92%). However, the highest levels of sensitivity and specificity were obtained in projects involving extracted viral RNA (Table 1). Most of the reviewed papers (8/10) related to RT-LAMP-CRISPR-Cas technology used samples treated with extraction kits. Moreover, only 1 study applied the Cas13 enzyme as an effector protein and used RNA extracted using a kit. In this case, the values for the calculated data were 73% to 97% for sensitivity, 95% to 100% for specificity, 90% to 100% for PPV, and 50% to 95% for NPV (Table 1).
The best results for collateral-based detection reaction were achieved with 50 nM Cas13a enzyme and a Cas13a/crRNA molar ratio of 2:1. On the other hand, the HybriDetect lateral flow kit showed higher sensitivity when reporter 2 was used at a final concentration of 1,000 nM and the assay buffer was supplemented with 5% polyethylene glycol (PEG). Determination of the limit of detection (LOD) of the CRISPR-Cas13a-based technology revealed that this technique detects as few as 1 to 10 SARS-CoV-2 particles (Fig. 2). After proteinase K-heat inactivation (PK-HID) treatment, the LAMP-CRISPR-Cas13a technique correctly detected samples with a cycle threshold (CT) value of <20 as positive. From samples with CT values of 20 to 30 and >30, the technique identified coronavirus as present in 83.3% and 62.5% of the samples, respectively (Fig. 3C). Finally, the CRISPR-Cas13a technology did not detect SARS-CoV-2 infection in negative samples (Fig. 3A). Based on the results obtained (Fig. 3B), we estimated that the RT-LAMP-CRISPR-Cas13a method for COVID-19 detection exhibits 100% specificity and 83% sensitivity, as well as a PPV of 100% and NPVs of 100%, 81%, 79.1%, and 66.7% for CT values of <20, 20 to 30, >30 and overall, respectively (Fig. 3C). The statistical analysis yielded a receiver operating characteristic (ROC) curve with an area under the curve (AUC) of 0.84 (95% confidence interval [CI], 0.73 to 0.93) (Fig. 4A); in addition, examination of the scatterplot revealed that diagnostic results could be confused in nasopharyngeal samples with a CT value of >30 (Fig. 4B).
Study of the state of the art revealed that greater efforts must be made to innovate in diagnostic methods; Bhatt et al. (20) found 1,286 papers related to RT-LAMP and CRISPR for SARS-CoV-2 diagnosis (surprisingly, only 98 of these applied RT-LAMP integrated with CRISPR-Cas technology), in contrast with the 7,000 studies involving RT-PCR. This indicates that efforts should also be focused on developing more efficient RT-LAMP-CRISPR-Cas protocols without RNA purification, which would reduce the cost of the testing and also produce results faster. Only 3 of the 20 papers reviewed did not use an RNA extraction kit (22–24). In addition, there are several advantages to the application of Cas13 endonuclease, as it has been reported to be more specific than other effector proteins (16). In this work, our research group developed an RT-LAMP-CRISPR-Cas13a protocol for diagnosing SARS-CoV-2 infection with an LOD of 10 viral copies, which is similar to the LOD of the RT-PCR method, considered the gold standard for diagnosis of COVID-19 (5, 7–10, 41). However, it has been reported that the RT-PCR for SARS-CoV-2 detection has a limited sensitivity of 45% to 60% (10), while the RT-LAMP-CRISPR-Cas13a technology increases this value significantly, up to 83%. As previously mentioned, the gold standard shows downsides in terms of costs, implementation, and time consumption (5, 8, 10) that are surpassed by the RT-LAMP-CRISPR-Cas13a technique. (i) RT-PCR requires a high-quality RNA extraction method, while our technology is applied on samples processed using the simple PK-HID protocol. (ii) The gold standard depends on expensive equipment and specialized personnel, which raise the price per reaction and difficulty of use outside of laboratories; by contrast, the RT-LAMP-CRISPR-Cas13a protocol eliminates the need for a thermocycler and sophisticated readout equipment, allowing easier implementation. (iii) The RT-PCR protocol takes at least 4 to 6 h, in contrast with the RT-LAMP-CRISPR-Cas13a method, which takes less than 2 h. For all these reasons, this RT-LAMP-CRISPR-Cas13a-based assay is proposed as a strong option to replace the current molecular gold standard diagnostic test. Furthermore, considering the criteria recommended by the WHO (42), this novel technique fulfills the three key features of accuracy, accessibility, and affordability. This is because on the one hand, it showed an accuracy [(true positive {TP} + true negative {TN})/total] of 87.2%, and on the other hand, it is both accessible and affordable. Comparing our results on sensitivity, specificity, PPV, and NPV with those obtained in previous studies, we found that the specificity and PPV values of the RT-LAMP-CRISPR-Cas13a technology were higher than those in 7 of the 10 RT-LAMP papers reviewed (25–31), and in one case, the sensitivity of this novel technique was even higher (32). Moreover, this technique showed higher sensitivity and NPV values than those in 2 of the 10 RT-LAMP-CRISPR papers reviewed which applied an RNA extraction kit to the clinical samples (21, 33). Among the others, 7 of 8 studies used DNA target-endonuclease effectors, and thus, a higher sensitivity could be obtained due to the intrinsic stability of DNA in contrast to that of RNA molecules. The lower sensitivity of the RT-LAMP-CRISPR-Cas13a protocol (83%) than that described in a previous study (97.4%) could be explained by the fact that the researchers used an RNA extraction kit (Direct-zol), so that the RNA was purified and concentrated, and also that the results were revealed by fluorescence (34). ROC analysis has become a popular method for evaluating the accuracy of medical diagnostic systems, as it provides accurate indices for the techniques tested that are not distorted by fluctuations caused by the use of arbitrarily chosen decision criteria or cutoff points (43). The AUC determines the inherent ability of the test to correctly identify a person as infected or not, where an AUC value of 0.5 indicates an absence of capacity for discrimination between infected and healthy populations, a value of 0.5 to 0.7 is related to unsatisfactory discrimination, and the discrimination power is acceptable when the AUC value is between 0.7 and 0.8, excellent for values contained in the range 0.8 to 0.9, and perfect when the AUC is close to 1 (44). The value of the area under the ROC curve, calculated by statistical analysis, validated our RT-LAMP-CRISPR-Cas13a technique as a reliable diagnostic method. Furthermore, the results shown in Fig. 4B indicate that this protocol provides less accurate diagnostics when viral loads are low. However, we should bear in mind that at this stage of infection, individuals present almost no risk of being contagious (45, 46). In summary, the high levels of specificity, sensitivity, PPV, and NPV obtained using this promising protocol working with RNA extraction kit-free samples place the LAMP-CRISPR-Cas13a technology at the forefront of rapid and specific diagnostic methods for infectious diseases. Thus, this technique could be established as a diagnostic tool for detecting other viral (papillomavirus [47, 48], Zika virus [49, 50], dengue virus [50], African swine fever virus [51], Ebola virus [52]) and bacterial (53, 54) (tuberculosis [55]) diseases, as previously done by other authors for infections such as those caused by multiresistant pathogens (56, 57). However, Cas13 detection methods should be optimized to enable direct diagnosis without prior amplification of nucleic acids.
A study of the state of the art was conducted with the aim of comparing the use of different novel diagnostic techniques. First, we conducted a search in PubMed with the keywords “RT-PCR diagnosis COVID-19” and compared the output with the number of publications on RT-LAMP and RT-LAMP-CRISPR strategies for COVID-19 diagnosis (20). Then, we collected data on the different sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) from 10 papers related to RT-LAMP and 10 papers on the RT-LAMP-CRISPR-Cas COVID-19 diagnostic technique (21–40). We used the results to calculate the parameters needed for the comparison.
The nucleocapsid gene (GenBank Gene ID: 43740575) of the SARS-CoV-2 virus was selected for study due to the fact that it shows a higher abundance of subgenomic mRNAs than other targets, which boosts the sensitivity of the diagnostic technique (58). Furthermore, the mutation rate found in this gene is lower than that in other targets, such as the spike gene and the ORF gene (59, 60). The target sequence was analyzed in silico with the aim of designing specific primers for amplification of a genetic region without any previously described mutation (N gene region, 12 to 213 bp [N2 gene]) (61). Three pairs of LAMP primers were designed using PrimerExplorer V5 software (F3-B3, FIP-BIP, and Floop-Bloop) to amplify the SARS-CoV-2 N2 gene. The FIP LAMP primer contained the T7 polymerase promoter in its sequences for the subsequent transcription step (Table 2). Two different RNA reporters (reporters 1 and 2) were used to reveal the results in order to select the one with the best signal. Both contained a single isomer derivative of fluorescein modification (FAM) at the 5′ extreme and a biotin molecule at the 3′ extreme (Table 2).
Clinical samples were supplied by the Microbiology Service of the Teresa Herrera Materno Infantil Hospital (A Coruña, Spain). The samples (n = 133) were obtained from nasopharyngeal swabs for SARS-CoV-2 detection (Table 3).
Ethical approval was granted by the Galicia Drug Research Ethics Committee (CEIm-G), and internal ethical approval was received by the Institute of Research A Coruña (INIBIC) from Coruña Hospital (CHUAC) (2020/207).
For sample processing, a proteinase K-heat inactivation (PK-HID) protocol was applied to samples from swabs stored in viral transport medium (Gibco) (62) as follows. Aliquots (95 μL) of samples were treated for 15 min at 55°C with 5 μL of proteinase K (10 mg/mL; stock), prepared at 1 mg/mL in a final volume of 100 μL, and heat-inactivated at 98°C for 5 min. Finally, the extracted RNA samples were stored at −80°C.
Amplification using the RT-LAMP (WarmStart LAMP kit [DNA and RNA]; NEB) reaction was performed following the manufacturer’s protocol. Briefly, RNA samples (5 μL) were added to a reaction mix containing 12.5 μL of WarmStart LAMP 2× master mix and 2.5 μL of 10× primer mix (FIP-BIP, 16 μM; F3-B3, 2 μM; Floop-Bloop, 4 μM; stock) adjusted to a final volume of 25 μL with dH2O. The reaction mixtures were incubated at 65°C for 1 h.
Each Cas13a-based detection reaction mixture was incubated at 37°C for 30 min with the following reaction components: 2 μL of 10× cleavage buffer (200 mM HEPES, 90 mM magnesium chloride, 600 mM sodium chloride), 0.5 μL of deoxynucleoside triphosphates (dNTPs) (HiScribe T7 quick high-yield RNA synthesis kit), 0.5 μL of T7 polymerase (HiScribe T7 quick high-yield RNA synthesis kit), 20 U RNase murine inhibitor (NEB), 0.15 μL Cas13a endonuclease (25 nM; MCLAB), 0.5 μL crRNA (50 nM; IDT), 2 μL reporter (1,000 nM; IDT), and 5 μL of a cDNA sample, adjusted to a final volume of 20 μL with dH2O. Different concentrations of Cas13a and crRNA (200, 100, and 50 nM) were tested, and two different enzyme/guide molar ratios were used (1:1 and 2:1).
Results were revealed using the HybriDetect lateral flow assay as described by the manufacturer (Milenia Biotec), with some modifications. Briefly, 20 μL of collateral-based detection product was mixed with 80 μL of assay buffer in a 96-well plate. Immediately, the gold extreme of the trip was submerged in the mix and held for 2 to 3 min. Following the manufacturer’s instructions, the reactive strips required calibration before application for management of an optimal RNA reporter concentration, and as mentioned, reporters 1 and 2 were tested. The results obtained using two different assay buffers were also compared: the kit assay buffer and the same supplemented with 5% polyethylene glycol (PEG). The results obtained, i.e., true positive (TP), false positive (FP), false negative (FN), and true negative (TN), were used to calculate the following parameters: sensitivity (TP/TP+FN), specificity (TN/TN+FP), PPV (TP/TP+FP), and NPV (TN/TN+FN).
For estimating the number of initial SARS-CoV-2 viral particles that the CRISPR-Cas13a technology was able to detect, we serially diluted (1:10) the RNA extracted using hospital equipment from two clinical samples with CT values of 20 and 25. Finally, 5-μL aliquots of each dilution were used for calculation of the limit of detection (LOD). Here, we applied an estimated correlation between the CT value and the viral load.
Statistical analysis was conducted using the GraphPad Prism9 program to construct a receiver operating characteristic (ROC) curve with a confidence interval of 95% (Wilson/Brown method) and to construct a scatterplot of two groups (false-negative and true-positive samples) against the CT value of each sample. | true | true | true |
PMC9604169 | Sujitha Jayaprakash,Mangala Hegde,Bandari BharathwajChetty,Sosmitha Girisa,Mohammed S. Alqahtani,Mohamed Abbas,Gautam Sethi,Ajaikumar B. Kunnumakkara | Unraveling the Potential Role of NEDD4-like E3 Ligases in Cancer | 16-10-2022 | cancer,ubiquitination,NEDD4,E3 ligases,tumor suppressor,oncogene,targeted therapy | Cancer is a deadly disease worldwide, with an anticipated 19.3 million new cases and 10.0 million deaths occurring in 2020 according to GLOBOCAN 2020. It is well established that carcinogenesis and cancer development are strongly linked to genetic changes and post-translational modifications (PTMs). An important PTM process, ubiquitination, regulates every aspect of cellular activity, and the crucial enzymes in the ubiquitination process are E3 ubiquitin ligases (E3s) that affect substrate specificity and must therefore be carefully regulated. A surfeit of studies suggests that, among the E3 ubiquitin ligases, neuronal precursor cell-expressed developmentally downregulated 4 (NEDD4)/NEDD4-like E3 ligases show key functions in cellular processes by controlling subsequent protein degradation and substrate ubiquitination. In addition, it was demonstrated that NEDD4 mainly acts as an oncogene in various cancers, but also plays a tumor-suppressive role in some cancers. In this review, to comprehend the proper function of NEDD4 in cancer development, we summarize its function, both its tumor-suppressive and oncogenic role, in multiple types of malignancies. Moreover, we briefly explain the role of NEDD4 in carcinogenesis and progression, including cell survival, cell proliferation, autophagy, cell migration, invasion, metastasis, epithelial-mesenchymal transition (EMT), chemoresistance, and multiple signaling pathways. In addition, we briefly explain the significance of NEDD4 as a possible target for cancer treatment. Therefore, we conclude that targeting NEDD4 as a therapeutic method for treating human tumors could be a practical possibility. | Unraveling the Potential Role of NEDD4-like E3 Ligases in Cancer
Cancer is a deadly disease worldwide, with an anticipated 19.3 million new cases and 10.0 million deaths occurring in 2020 according to GLOBOCAN 2020. It is well established that carcinogenesis and cancer development are strongly linked to genetic changes and post-translational modifications (PTMs). An important PTM process, ubiquitination, regulates every aspect of cellular activity, and the crucial enzymes in the ubiquitination process are E3 ubiquitin ligases (E3s) that affect substrate specificity and must therefore be carefully regulated. A surfeit of studies suggests that, among the E3 ubiquitin ligases, neuronal precursor cell-expressed developmentally downregulated 4 (NEDD4)/NEDD4-like E3 ligases show key functions in cellular processes by controlling subsequent protein degradation and substrate ubiquitination. In addition, it was demonstrated that NEDD4 mainly acts as an oncogene in various cancers, but also plays a tumor-suppressive role in some cancers. In this review, to comprehend the proper function of NEDD4 in cancer development, we summarize its function, both its tumor-suppressive and oncogenic role, in multiple types of malignancies. Moreover, we briefly explain the role of NEDD4 in carcinogenesis and progression, including cell survival, cell proliferation, autophagy, cell migration, invasion, metastasis, epithelial-mesenchymal transition (EMT), chemoresistance, and multiple signaling pathways. In addition, we briefly explain the significance of NEDD4 as a possible target for cancer treatment. Therefore, we conclude that targeting NEDD4 as a therapeutic method for treating human tumors could be a practical possibility.
Cancer is one of the top causes of death, which accounts for 19,292,789 new cases and 9,958,133 deaths worldwide according to the GLOBOCAN 2020 [1,2,3]. It is a deadly disease, and it is believed that human beings have been inflicted by cancer for more than 200 million years [4]. Among various cancers, prostate and breast cancer (BC) account for a significant portion of cancer cases in men and women, respectively, and brain and hematological malignancies account for the highest percentage of cancer cases among children [1,5,6,7,8]. Importantly, cancer is generated by a series of gene mutations that alter the functions of cells, and these include both oncogenes and tumor suppressor genes [1,4,9,10,11,12,13,14,15,16,17]. Several studies suggest tumorigenesis as a multistep process involving genetic changes which cause typical cells to undergo a gradual change into extremely lethal variants [18,19,20]. Interestingly, several biomarkers have been discovered to have diagnostic and therapeutic utility, and globally increasing cancer incidence and mortality bring about a need for precise biomarkers for better detection, diagnosis, prognosis, and monitoring [11,21,22,23,24]. Currently, only a few cancer biomarkers are advised for clinical use, with the majority of them being used to assess treatment response in advanced cancer patients. However, the majority of cancer biomarkers now being used in clinical settings are less effective for mass screening or early diagnosis, and only a few new and helpful cancer biomarkers are being discovered and proven for screening and early diagnosis [24]. As a result, the discovery of novel biomarkers is needed for the identification and detection of cancer. The ubiquitin-proteasome system is an essential form of post-translational protein modification (PTM) that is important in protein homeostasis, as well as in the regulation of vital physiological activities [25,26,27]. Ubiquitination is extremely important in biological processes and alters post-translationally modified proteins in a way that causes the degradation, stabilization, or relocation of substrates [27,28,29,30,31,32]. The ubiquitin-mediated proteolytic pathway is a highly adaptable and reversible process mediated by an enzyme cascade [27]. Ubiquitination regulates thousands of intracellular protein levels in cells and plays a role in practically all cell physiology and disease, including DNA damage and repair, cell death, and immunological responses, by initiating selective proteolysis via the 26S proteasome [27,33,34]. Ubiquitin (Ub) is a highly conserved regulatory protein with 76 amino acids that can covalently tag target proteins through a series of enzymatic processes involving the enzymes Ub-activating (E1), Ub-conjugating (E2), and Ub-ligating (E3) [35,36,37,38]. Deubiquitinases (DUBs), which are also essential for nearly all cellular signaling pathways, including the cell cycle, apoptosis, receptor downregulation, and gene transcription, can similarly reverse the activity of Ub ligases by removing Ub from substrate proteins [39,40]. Ubiquitination is activated via a sequential cascade including E1, E2, and E3, with E1 first triggering Ub via an ATP-dependent process [35,41]. The activated Ub is then transported to an E2 catalytic cysteine residue and finally binds to the specific target (Figure 1) [42,43,44]. E3 ligases play the most important role in recognizing the target protein and regulating the covalent connection between the target and ubiquitin moieties in this enzyme cascade [44]. The human genome already has 600 E3s, which are mostly divided into three groups: E3s from the RING finger family, the RING-between-RING (RBR) family, and the E6-AP C-terminus (HECT) family [45]. The E3 subfamily’s pivotal members are HECT-type E3s, which have a typical C-terminal module containing 350 amino acids [46,47,48]. These E3s are classified into three groups, HERC (HECT and RCC-like domain) E3s, neural precursor cell-expressed developmentally downregulated 4 (NEDD4)/NEDD4-like E3s, and other E3s [46,47,48]. Numerous studies have reported that tumor formation and incidence have been linked to abnormalities in ubiquitination [40,49,50]. As E3s are known to have a significant role in the system’s substrate specificity, their abnormalities might contribute to the development of human cancer [51,52]. Moreover, ubiquitination tightly regulates components in both tumor-suppressing and tumor-promoting pathways [53,54]. There have been many targeted therapies developed to battle carcinogenesis based on altered components such as the E3 ligases, E1, E2, deubiquitinases (DUBs), and proteasomes [27,55,56]. Among those, E3 ubiquitin ligases (E3s) are crucial ubiquitination enzymes and play a vital role in cancer development [53,57]. Moreover, ubiquitin ligase mutations and changes are seen in a diverse spectrum of tumors and have a significant impact on clinical outcomes [53,58,59]. In several studies, it was reported that among the HECT family E3s, the function of NEDD4/NEDD4-like E3 ligases influences cancer cell proliferation, migration, and invasion, as well as anticancer therapy sensitivity, via regulating several substrates [53,54,60]. The NEDD4/NEDD4-like E3 ligase is a ubiquitin–protein ligase that could regulate a variety of membrane proteins to help them internalize and turnover [61]. It also has a significant impact on central nervous system development, hypertension regulation, etc. [45]. The NEDD4 protein is mostly found in the cytoplasm, particularly near the nucleus, and in humans the NEDD4 gene is located on chromosome 15q21.3 and has 33 exons that code for 120 kDa protein [54,62,63]. The NEDD4 structure has a C-terminal HECT domain for ubiquitin–protein ligation, an N-terminal C2 domain for membrane binding, and a central two to four double tryptophan residue (WW) domain for protein–protein interaction (Figure 2) [64]. The C2 domain is a 116-amino-acid-long calcium-dependent lipid-binding region that directs proteins to phospholipid membranes and functions in protein–protein interactions. WW domains regulate the protein–protein interactions that are typically 40 amino acids long and hold two conserved tryptophan (W) residues separated by 21 amino acids. A conserved cysteine residue in the HECT domain creates an intermediary thioester bond with active ubiquitin received from an E2 before catalyzing lysine ubiquitination in the substrate protein [64,65,66,67]. In mammals, the NEDD4 family consists of nine members: NEDD4, NEDD4 Like E3 Ubiquitin Protein Ligase (NEDD4L) also known as NEDD4-2, the SMAD Ubiquitylation Regulatory Factors 1 (SMURF1) and 2 (SMURF2), WW Domain Containing E3 Ubiquitin Protein Ligase 1 (WWP1) and 2 (WWP2), NEDD4-like ubiquitin–protein ligase 1 (NEDL1) and 2 (NEDL2) and the Itchy E3 Ubiquitin–Protein Ligase Homolog protein (ITCH) (Figure 3) [62]. All of these members of the NEDD4 subfamily share a common functional domain structure (Figure 4) [68,69]. The NEDD4 subfamily members are involved in a variety of cellular processes. The proteins NEDD4 and NEDD4-2 were the first members of this subfamily to be found and are now the most investigated. The subfamily members are considered to be involved in a variety of illnesses, including cancer and neurological diseases [60,70]. NEDD4-2 functions in a variety of ion channels, including chloride, potassium, and sodium [71,72,73]. It also interacts with proteins involved in the Wnt signaling pathway, the TGF-β signaling pathway, and the autophagy process [74,75,76]. ITCH regulates a wide range of biological mechanisms because of a large number of target proteins [77]. In addition, ITCH regulates the TGF-β signaling pathway and influences cancer [78]. Another member, WWP1, is a versatile protein that has many targets including ErbB4/HER4, JunB, Smad2, Smad4, and p53. As a result, WWP1 is involved in cancer, protein degradation, protein trafficking, transcription, viral budding, and infectious and neurological diseases [79]. WWP2 binds to substrates in many signaling pathways, including the PI3K/Akt and TGF-β pathways, and is linked to cancer and immune system modification [80]. Both SMURF1 and SMURF2 suppress the TGF-β/BMP signaling pathways [81,82]. SMURF1 is involved in the noncanonical Wnt and MAPK pathways and regulates cell growth and morphogenesis, cell migration and polarity, and autophagy [83]. SMURF2 has been linked to similar pathways. Its dual role in cancer is frequently studied, as it can serve both as a tumor suppressor and an oncoprotein [83]. NEDL1 and NEDL2 are also known as the HECT, C2, and WW domains containing E3 ubiquitin–protein ligase 1 (HECW1) and 2 (HECW2) [70]. By ubiquitinating and degrading Dishevelled-1(Dvl1), NEDL1 participates in the Wnt signaling pathway [84,85]. Recent evidence suggests that NEDL1 is also involved in the TGF-β signaling pathway via Smad4 ubiquitination. These two proteins, HECW1 and HECW2, appear to interfere with a variety of physiological functions, including the enteric nervous system and kidney development [86,87]. NEDD4 mediates receptor-mediated endocytosis along with proteasomal degradation and ubiquitination of its substrate proteins [63]. NEDD4 E3 ligases modulate signaling molecule trafficking by monoubiquitination and K63-linked polyubiquitination and hence play a critical role in cellular activities [63]. The NEDD4 is an important regulator of various cellular physiological processes such as endocytosis, cell growth, neuromuscular junctions, and modulation of epithelial Na+ channels (ENaC) [92,93,94]. In addition, NEDD4 is also linked to several diseases, including cancer, cardiovascular disease, metabolic disease, neurological disorders, renal disease, and others. For example, reduced NEDD4 function causes significant heart arrhythmia by changing cardiac ion-channels following transcription [95]. Further, adult mice with conditional NEDD4L deletion in lung epithelial cells were shown to develop chronic lung illness with increasing fibrosis and bronchiolization. These mice also showed increased Muc5b expression in peripheral airways, honeycombing, and distinctive changes in the lung proteome [96]. Another in vivo study found that NEDD4-2 loss causes an unanticipated progressive kidney damage phenotype that includes increased Na+ reabsorption, increased Na+Cl− and ENaC cotransporter expression, hypertension, and low aldosterone levels in animal models [71].
NEDD4 has a significant role in the development and progression of various types of cancers by targeting different substrates. Interestingly, it is found that there are many genes upstream and downstream of NEDD4, and it has the potential to be employed as a molecular switch to control tumor development via these competitive substrates [54]. Expectedly, NEDD4 has been associated with the regulation of several signaling pathways. For example, the WW domain of NEDD4 contributes to polyubiquitination and degradation of Smad2/3, limiting TGF-β signaling by specifically recognizing the TGF-β-induced phospho Thr-ProTyr motif in the junction region [74]. Further, NEDD4 causes the ubiquitination of Unc-51 like autophagy activating kinase 1 (ULK1), a protein involved in the initiation of autophagy, and controls the process [75]. According to several studies, NEDD4 acts as an oncoprotein that promotes cancer cell development. For example, a study demonstrated that siRNA reduction of NEDD4 resulted in a decrease in cell proliferation and changed cell shape in LoVo and HCT-15 colon cancer cells. According to Amodio et al., NEDD4 overexpression was found in non-small-cell lung cancer (NSCLC) cell lines, while NEDD4 knockdown in vitro and in vivo dramatically decreased NSCLC cell proliferation and tumor growth, respectively. NEDD4’s carcinogenic effect in NSCLC cells can be due to the inactivation of PTEN [97]. NEDD4 is the key pro-oncogenic ligase that mediates the ubiquitination of PTEN; however, NEDD4L is mainly involved in the degradation of the tumor suppressor gene p53 [98,99,100]. Moreover, in breast and prostate cancer cells, NEDD4 might enhance cell proliferation and migration by regulating the PTEN/Akt signaling pathway [101]. However, Eide et al. reported that NEDD4 increased colon cancer cell proliferation without relying on the PTEN or PI3K/Akt signaling pathways [102]. A study reported that NEDD4 targets Myc oncoproteins, c-Myc and N-Myc, for degradation by directly binding to them, which resulted in downregulation of these proteins and suppression of cell proliferation in pancreatic cancer and neuroblastoma [63,103]. Several NEDD4 substrates, such as Ras, Notch, pAkt, VEGFR2, Mdm2, and FGFR, have been identified as oncoproteins in carcinogenesis, suggesting that using NEDD4 to target their breakdown could have anticancer effects [63]. A plethora of studies have shown that NEDD4 promotes tumor growth in diverse types of cancers, and these studies suggest that the role of NEDD4 in tumor progression is context-dependent [60,63]. This pleiotropic nature of NEDD4 is mainly attributed to its ubiquitous expression in a wide range of tissues. NEDD4 proteins have several downstream target substrates that arbitrate various functionalities, and numerous receptors that share the same protein trafficking and endocytic machinery are controlled by various NEDD4-like E3s [60]. Therefore, in malignancies, NEDD4 behaves as both an oncogene and a tumor suppressor, thus NEDD4 inhibitors or activators are needed for cancer treatments [54,63]. NEDD4 was found to be an oncogene because of its role in inhibiting PTEN, a well-known tumor suppressor [104]. In several types of human cancer cell lines, PTEN degradation and increased NEDD4 levels have been discovered [104]. PTEN is a phosphatase that inhibits tumor growth in a variety of malignancies by suppressing the PI3K/Akt signaling pathway, which is required for the survival of cancer cells. NEDD4 has been identified as a PTEN-specific E3 ligase and has been shown to promote Akt signaling by lowering the amount of PTEN protein. Thus, NEDD4 has been proposed as an oncoprotein and a possible drug target [60]. Aforementioned studies highlight NEDD4 as a potential biomarker for poor prognosis as well as a therapeutic target for the treatment of different cancer types. In the current article, we discuss the function of the NEDD4 protein in the development and suppression of cancer. Then, we describe the vital roles of NEDD4 in carcinogenesis and development, which include cell survival, cell proliferation, autophagy, cell migration, invasion, metastasis, epithelial–mesenchymal transition (EMT), chemoresistance, multiple signaling pathways, oncogenic role, and tumor-suppressive role (Figure 5). These studies suggest that NEDD4 can be used as a drug target for human cancer; therefore, we briefly discuss the clinical significance of NEDD4 in this article.
NEDD4 is known to induce cancer cell survival in different cancers. For example, NEDD4 overexpression leads to the suppression of PTEN expression and induces Notch-1, which leads to decreased apoptosis in RT4 bladder cancer (BCa) cells. Moreover, this study also showed that suppression of NEDD4 leads to BCa cell apoptosis [105]. Similarly, in QGY7703 and SMMC7721 liver cancer cells, NEDD4 acts as an oncoprotein, and overexpression of this protein was shown to suppress apoptosis via partially increasing large tumor suppressor kinase 1 (LATS1) expression [106]. Moreover, LATS1/2 is ubiquitinated by NEDD4 when it recognizes the PY motif in proteins [107]. In addition, NEDD4 downregulation resulted in DU145 cell apoptosis via both death receptor and mitochondria-dependent mechanisms [108]. A recent study showed that overexpression of FOXA1 reduces the proliferation induced by NEDD4 upregulation and promotes apoptosis in Caco-2/LoVo colon cancer cells [109]. In conclusion, NEDD4 has a crucial role in the survival of cancer cells.
The function of NEDD4 in cancer cell proliferation is demonstrated by several studies. For example, the overexpression of NEDD4 increased the proliferation in HeLa cells [108]. Similarly, in QGY7703 and SMMC7721 liver cancer cells, the overexpression of NEDD4 suppressed LATS1, which results in enhanced cell proliferation [106]. Further, NEDD4 knockdown inhibits cell proliferation in Huh-7 HCC cells through overexpression of PTEN and consequent inactivation of Akt, ERK1/2, and STAT3 [110]. A similar study also revealed that in Huh7 HCC cells, depletion of NEDD4 has a detrimental effect on cell proliferation. NEDD4 appears to be important in promoting HCC proliferation and spread through the stimulation of the PTEN/PI3K/Akt signaling pathway [111]. In PANC-1 pancreatic ductal adenocarcinoma (PDAC) cells, NEDD4 depletion suppresses cell growth partially by regulating the PTEN/Akt signaling pathway [112]. In addition, downregulation of NEDD4 suppresses proliferation of RT4 BCa cells. Inhibition of NEDD4 activated PTEN while suppressing Notch-1, indicating that NEDD4’s oncogenic activity was mediated partially by downregulation of PTEN and overexpression of Notch-1 [105]. Interestingly, another study suggested that NEDD4 can inhibit phosphatidylinositol 4-phosphate 5-kinase α (PIP5Kα)-dependent phosphatidylinositol 4,5-bisphosphate (PIP2), leading to a significant proliferation defect [113]. In conclusion, these studies have established that NEDD4 can affect proliferation in various cancer cells.
Autophagy is an evolutionarily conserved catabolic mechanism that eukaryotic cells use to degrade or recycle internal contents through a membrane-trafficking pathway [114,115]. In distinct phases of cancer development, autophagy performs a dynamic tumor-promoting or tumor-suppressing role [116,117]. It has shown that autophagy was suppressed in lung and prostate carcinoma cell lines when NEDD4 was knocked out and this NEDD4 downregulation significantly increased activated mTOR (p-mTOR) levels, implying that mTOR signaling was involved in NEDD4-mediated autophagy [108]. Another study has shown that high levels of advanced glycation end product (AGE) levels increased NF-κB expression in Hep3B and Saos-2 cell lines, which is amplified more in the absence of p53. The NF-κB-dependent expression of NEDD4 enhances Beclin 1 cleavage, impairing autophagy and causing autophagosome build-up, which leads to cell death [118]. Beclin 1 is a subunit of the PI3K-III complex and a tumor suppressor that plays an important function in autophagy and apoptosis. This protein has a sequence that looks like a PY motif (LPxY), and the WW domains of NEDD4 interact with it specifically. NEDD4 was shown to polyubiquitinate Beclin 1 and regulate its stability, which suppresses autophagy [60,119]. Further, in pancreatic cancer cells, NEDD4 prevents autophagy activity under metabolic stress by decreasing mitochondrial function and limiting tumor development by destabilizing an autophagy protein, ULK1, and a glutamine transporter, ASCT2 [75]. Taken together, NEDD4 seems to prevent autophagy in cancer cells.
According to several studies, it was noted that NEDD4 overexpression increases cell migration and invasion in various cancers. For example, a study demonstrated that NEDD4 overexpression decreased PTEN and increased Notch-1 levels, resulting in invasion and migration of RT4 BCa cells [105]. Another study in colon cancer tissues demonstrated that NEDD4 promotes the progression of cancer by activating FOXA1 ubiquitination in conjunction with the inhibition of miRNA-340-5p and the overexpression of ATF1 [109]. Similarly, downregulation of NEDD4 using curcumin in pancreatic cancer cell lines Patu8988 and Panc-1 resulted in a considerable reduction in tumor cell development, in association with increased expression of PTEN and p73 [120]. Further, in QGY7703 and SMMC7721 hepatocellular carcinoma (HCC) cells, NEDD4 increased cell migration and invasion via inhibiting the tumor suppressor LATS1 [106]. Another study reported that NEDD4 is closely linked to the development and progression of lung cancer due to its overexpression in a significant fraction of NSCLC, promotion of PTEN degradation, and enhancement of the malignant characteristics of lung epithelial cells [97]. Similarly, an in vivo study in NSCLC showed that silencing of NEDD4 inhibited cell invasion and migration by increasing PTEN expression and by inhibiting PI3K/Akt signaling pathway [121]. Moreover, NEDD4 knockdown in A549 NSCLC cells prevented EGF-induced cell migration through regulating its interaction with the EGFR signaling complex and cathepsin B lysosomal secretion [122,123]. According to another study, the influence of proline-rich γ-carboxyglutamic acid protein 4 (PRRG4) on invasion and migration in BC cells is mediated via both NEDD4 binding and Robo1 downregulation [124]. In U251 glioma cells, it was reported that NEDD4 overexpression enhanced the cell invasion and migration due to ubiquitination and degradation of cyclic nucleotide ras guanine nucleotide exchange factor (CNrasGEF) [60]. Moreover, NEDD4 promoted the ubiquitination of NEDD4/Rap2a, which eased the invasion and migration of U251 and U87 cancer cells. In addition, human glioma cells showed suppression of invasion and migration when NEDD4 was downregulated, but overexpression of NEDD4 reversed these effects. These findings imply that NEDD4 has a significant role in glioma development [125]. It was also shown that NEDD4 mediates EGF-dependent AGS and N87 gastric cancer cell line invasion and migration via EGFR-mediated metastasis signaling [126]. Taken together, NEDD4 increases cell migration and invasion in human cancer cells.
Metastasis is the process by which cancer cells spread from the primary tumor to surrounding tissues and distant organs, and it is the leading cause of cancer morbidity and mortality [127,128,129]. In several studies, NEDD4 has been described to take part in the progression of metastasis. For example, the stability of NEDD4 is regulated by p34SEI-1, a 34 kDa oncoprotein that controls PTEN degradation and induces the PI3K/Akt pathway, which results in tumorigenesis and metastasis [130]. Another study revealed that PRRG4-mediated recruitment of NEDD4 stimulates ubiquitination and degradation of Robo1 (a tumor suppressor gene), subsequently leading to the activation of protein tyrosine kinases Src and FAK, which are critical for BC cell motility, invasion, and metastasis [124]. An in vivo study demonstrated that overexpression of NEDD4 in CC tissues and Caco-2/LoVo cell lines aided in the development of xenograft tumors and metastasis, as well as tumorigenesis in mice models [109]. Interestingly, another study showed NEDD4 to be important in gastric cardia adenocarcinoma (GCA) metastasis, as the shRNA-mediated NEDD4 decrease in AGS and N87 gastric cancer cells impairs basal and EGF-stimulated cell invasion and migration [126]. It was also shown that NEDD4 is important for lung cancer metastasis by increasing cathepsin B secretion or causing PTEN degradation. However, NEDD4-2 has more complex functions than NEDD4 as it inhibits lung cancer cell metastasis; on the other hand, it enhances lung cancer survival by causing general control nonderepressible 2 (GCN2) degradation [123]. Further, NEDD4 appears to play a significant role in promoting HCC metastasis via stimulating the PTEN/PI3K/Akt signaling pathway [111].
The epithelial–mesenchymal transition (EMT) is a process that can be reversed and temporarily transforms epithelial cells into quasi-mesenchymal cell types, and this process plays a significant role in chemoresistance and metastasis [131,132,133,134,135,136]. In CNE1 and CNE2 nasopharyngeal carcinoma cells, it was found that partial overexpression of the NEDD4 signaling pathway caused EMT in DDP-resistant cells. Moreover, EMT shifts to MET when NEDD4 is deleted via an shRNA, making DDP-resistant cells susceptible to DDP [137]. NEDD4 has been linked to lysosomal-associated protein transmembrane 5 (LAPTM5), and the downregulation of LAPTM5 in T24 and 5637 BCa cells suppressed proliferation and hindered invasion and migration through the inhibition of EMT markers, which hinted that NEDD4 can be involved in LAPTM5-influenced EMT [138]. Another study showed that miR-93 overexpression in Huh-7 cancer cells facilitated TGF-β-induced EMT by inhibiting NEDD4L expression [139]. Notably, NEDD4 overexpression was observed to induce the EMT, which might be responsible for its role in invasion and metastases.
Chemoresistance is often cited as a primary cause of cancer therapeutic failure, resulting in cancer relapse and spread [140,141,142]. NEDD4 has shown to play a major role in chemoresistance in various cancers. A study on lung cancer proved that upregulation of NEDD4 via the PI3K/Akt pathway might contribute to tumor growth and modulate lung ADC chemoresistance [121]. As previously mentioned, NEDD4 is carcinogenic in CNE1 and CNE2 NPC cells because it increases the EMT features of DDP-resistant cells. It was also discovered that knocking down NEDD4 with shRNA reversed EMT and sensitized DDP-resistant cells. These data suggest that NEDD4 is involved in chemoresistance in NPC cancer cells [137]. Moreover, a study on U87MG and U251 GBM cell lines demonstrated the crucial role of NEDD4 in controlling the redox imbalance in temozolomide (TMZ)-resistant GBM cells by PTEN degradation and activation of the Akt/NRF2/HO-1 signaling pathway [143]. Another similar study in GBM revealed that glioma-linked oncogenic lncRNA LINC01198 promotes proliferation of gliomas and temozolomide resistance by acting as a scaffold and enlisting NEDD4 enzymes to attack certain proteins such as PTEN [144]. In A375 melanoma cell lines, erastin, a ferroptosis activator, stimulates NEDD4 and FOXM1 expression, and it led to the ubiquitination and degradation of voltage-dependent anion channels, VDAC2/3, which further led to the suppression of erastin-induced ferroptosis in melanoma [145]. Another study in H1650/ER cells showed that NEDD4 might enhance NSCLC-acquired resistance to erlotinib by lowering the expression of PTEN [146].
NEDD4 is a protein that regulates many signaling pathways. Wnt/β-catenin signaling is negatively regulated by NEDD4, which targets the LGR5 receptor and DVL2 for proteasomal and lysosomal degradation. Moreover, inactivation of NEDD4 and NEDD4L promotes Wnt activation, which improves tumor propensity and progression [147]. On the other hand, knockdown of NEDD4 decreases pAkt levels, increases PTEN, and inhibits the development and migration of the Hep3B HCC cell line [60,111]. According to another study, NEDD4 promotes insulin-like growth factor-1 (IGF-1) signaling and mitogenic activity by interacting and conjugating monoubiquitin to IRS-2, improving IRS-2-mediated signaling and cell proliferation triggered by IGF-1. Interestingly, in PC-3 prostate cancer cells, the NEDD4 and IRS-2 connection is also essential for optimal IGF-1 signaling activation and cell proliferation [148]. Another study showed that NEDD4 ubiquitylates and destabilizes WW45 kinase and LATS1/2, which are needed for active Hippo signaling. NEDD4 positively regulates cell proliferation through the negative regulation of LATS1/2 and WW45. This suggests that NEDD4 influences cell proliferation and cancer growth through Hippo signaling. As a result, NEDD4 functions as a regulator in the Hippo pathway, and the depletion of NEDD4 causes an increase in apoptosis and a decrease in proliferation [149].
A surfeit of studies suggests that the NEDD4 protein has both oncogenic and tumor-suppressive properties. Several preclinical studies revealed that NEDD4 has cancer chemopreventive and therapeutic capabilities against various malignancies, which include solid tumors such as brain, breast, colon, bladder, liver, pancreas, prostate cancers, etc., and hematological cancers such as lymphoma and leukemia. Table 1 and Table 2 summarize the role of the NEDD4 protein in various types of cancers. These studies, which are briefly discussed below, reveal that the NEDD4 protein has immense potential as a successful target for both the prevention and treatment of many malignancies.
Bladder cancer (BCa) is one of the most frequently occurring malignancies of the urinary bladder, which accounts for 3 percentage of global cancer diagnoses according to GLOBOCAN 2020 [105,138,200,201,202]. Interestingly, it was noted that downregulation of NEDD4 resulted in the inhibition of cell proliferation, invasion, and migration in RT4 BCa cells. Moreover, NEDD4 inhibition activated PTEN while suppressing Notch-1. Therefore, downregulation of NEDD4 in BCa cells reduced cell proliferation, induced apoptosis, and hindered cell migration [105]. This study pointed out the importance of the development of NEDD4 inhibitors for the treatment of human BCa.
BC is the most frequent type of cancer in women all over the world, and triple-negative breast cancer (TNBC) is characterized as a very aggressive malignancy with a bad prognosis and no viable targeted therapy available for its treatment [203,204,205,206,207,208,209,210]. NEDD4’s role in the development of BC has been reported in several studies [54,104,178,211]. It has been observed that NEDD4 is predominantly overexpressed in HER2-amplified BC [212]. The expression of NEDD4L was also reported to be associated with a better prognosis for BC survivors [213]. A study revealed p34SEI-1 as a highly expressed oncoprotein in human BC tissues which causes tumorigenesis by triggering NEDD4 which in turn causes poly-ubiquitination and degradation of PTEN, thereby promoting tumorigenesis by favorably regulating the PI3K/Akt pathway [151]. According to another study, PRRG4’s influence on migration and invasion in BC cells is mediated via both NEDD4 binding and Robo1 downregulation [124]. Another in vitro study suggests that using the MDA-MB-231 cell line, Cdh1 suppressed the E3 ligase activity of WWP2 in the MDA-MB-231 cell line in a manner that is distinct from the anaphase-promoting complex and cyclosome. Therefore, the lack of Cdh1 activates WWP2, which in turn causes a decrease in the abundance of WWP2 substrates such as PTEN, which stimulates PI3K/Akt oncogenic signaling and promotes cancer growth [173]. Another study in the MDA-MB-231 cell line revealed that the expression of the tumor suppressor Mig6 is stabilized by type I γ phosphatidylinositol phosphate 5-kinase i5 (PIPKIγi5). Mig6 expression is lost as a result of PIPKIγi5 knockdown, which significantly increases and prolongs EGFR-mediated cell signaling. Direct interaction between PIPKIγi5 and NEDD4 disrupts the process of Mig6 ubiquitination and subsequent proteasomal destruction. Therefore, in PIPKIγi5-knockdown cells, a lack of NEDD4 can restore the expression of Mig6 [174]. NEDD4 was more likely to bind to connexin 43 (Cx43) when ER was inhibited, which caused Cx43 to be ubiquitinated. Additionally, tamoxifen and fulvestrant suppressed ER and phosphorylated the MAPK subunit p38, and downregulated Rac and MKK3/6. Further, pretreatment with Akt and MAPK inhibitors reversed fulvestrant-reduced Cx43 expression. These results demonstrate that in ER-dependent ER-positive BC cells and Cx43 expression might positively influence cell motility [177]. ARRDC3 (arrestin domain-containing 3) is a metastasis suppressor that blocks EGF-driven endocytic recycling of ITGβ4 by promoting NEDD4-dependent ubiquitination. By lowering ITGβ4 levels in EVs, ARRDC3 lowers the metastatic capacity of BC-cell-derived extracellular vesicles (EVs) [175]. Histopathological evaluation has shown that WWP1, as well as NEDD4 expression, were modulated in the BC tissues [150,152]. Moreover, knockdown of WWP1 in MCF-7 and T47D BC cells has shown a decrease in cell growth and colony formation. Hence, WWP1 might function as a co-activator or anti-apoptotic factor to encourage the survival and growth of BC cells [150]. Another study in MCF-7 and T47D cell lines reported that WWP1 negatively regulates LATS1, a tumor suppressor, through its degradation by polyubiquitination and the 26S proteasome pathway. Moreover, it was reported that the knockdown of NEDD4 and ITCH increased the level of LATS1, whereas the knockdown of SMURF1 decreased the level of LATS1 [171]. In addition, robustaflavone-A (RF-A) dramatically increased MCF-7 ferroptosis, resulting in lipid peroxidation and ROS generation by increasing the expression of VDAC2 channels and decreasing the expression of NEDD4 [180]. Apart from the oncogenic property of NEDD4 revealed in several studies, some reports state the tumor-suppressing role of NEDD4 in BC. For example, a study has demonstrated the tumor-suppressing role of NEDD4 by which PIP5Kα, a ubiquitinated protein that enhances the proliferation, invasion, and migration of BC cells, was impeded by NEDD4. PIP5Kα produces PIP2, a key regulator of lipids in a variety of physiological activities. The oncogenic PI3K/Akt pathway can also act upstream via PIP5K-dependent PIP2 production. PIP5Ka was ubiquitinated and proteasomally degraded by NEDD4, resulting in a decrease in plasma membrane PIP2. Furthermore, deletion of the PIP5Ka gene reduced EGF-induced Akt activation and produced a major proliferation deficit [113]. In another study, the knockdown of NEDD4 showed to increase the levels of MAPK phosphatase 3 (MKP3) in MDA-MB-231 BC cells [170]. Another study demonstrated a negative correlation between HER3, a key player in cancer, and NEDD4 levels. More significantly, the NEDD4 knockdown increased HER3 expression, sensitizing cancer cells to the anti-HER3 antibody’s ability to suppress cell proliferation. Collectively, these findings imply that low NEDD4 levels might be able to anticipate when HER3 signaling is activated. In addition, NEDD4 knockdown along with NRG-1 and HER3 mAb in an MCF-7 (shNEDD4) mice xenograft showed decreased tumor volume [172]. Further, an in vitro study showed decreased ER and HER3 expression and increased proliferation rate in shNEDD4 knockdown MCF-7 cells [176]. Another in vitro study demonstrated that SMURF1 knockdown increased HER2 expression in BT474 cells [179]. In summary, NEDD4-like E3 ligases were found to be a feasible therapeutic target as well as a potential prognostic predictor for BC.
Cervical cancer is the fourth most prevalent cancer in women worldwide and it is the most common gynecological malignancy in developing nations, and has an incidence rate of less than 1 percentage and mortality of less than 0.5 percentage [214,215,216,217,218,219]. A study in HeLa cells reported that NEDD4 prevented PTEN from inducing apoptotic cell death, and PTEN, in turn, decreased the level of NEDD4 [181]. Another study in HeLa cells described NEDD4 as a novel binding partner of Beclin 1, and this study showed that NEDD4 polyubiquitinates Beclin 1 with Lys11 and Lys63 linked chains. Importantly, NEDD4 expression regulates Beclin 1 stability, and it is degraded by the proteasome via Lys11-linked polyubiquitin chains when the Beclin 1 interacting protein VPS34 is depleted. Therefore, Beclin 1 is the first tumor suppressor to be described as being under the control of Lys11-linked polyubiquitination [119]. Further, NEDD4L overexpression specifically downregulates ULK1 protein levels by preferentially ubiquitylating it for degradation by the proteasome, which actively transcribes ULK1 mRNA. Basal levels of ULK1 are immediately restored upon reactivation of mTOR-dependent protein synthesis; however, mTOR inhibits the function of newly synthesized ULK1. This gets the cell ready for another round of potential autophagy stimulation in HeLa cells [182]. Furthermore, a study in HeLa cells demonstrated that NEDD4 controls the level of Cx43, which acts as a tumor suppressor both at basal conditions and in response to protein kinase C activation. Additionally, it was discovered that NEDD4 expression led to the total loss of gap junctions and an increase in the lysosomal degradation of Cx43 in HeLa cancer cells [93].
Colorectal cancer is one of the most common cancers in women and men worldwide. Both hereditary and environmental factors play a key role in colorectal cancer etiology [220,221,222]. A study revealed that NEDD4 was found to be expressed in 80 percentage of colorectal carcinomas and has tumorigenic activity by inducing ubiquitination and degradation of the PTEN gene [153]. Another study on HCT-15 and LoVo colon cancer cells observed that PTEN and PI3K/Akt signaling are not required for NEDD4 to induce colon cancer cell growth. Moreover, NEDD4 knockdown induced alterations in cell morphology and reorganization of the actin cytoskeleton [102]. Another study demonstrated that the expression of NEDD4 has a proportional effect on cell growth and metastasis by varying vimentin, N-cadherin, snail, and ATF-1 expression in LoVo and Caco-2 cells. Further, in LoVo cells, the overexpression of NEDD4 reduced the expression of E-cadherin, FOXA1, and miR-340. Furthermore, in Caco-2 cells, overexpression of NEDD4 increased cell proliferation and colony number and reduced apoptosis by decreasing Cyto-C, PUMA, Apaf-1, Bax, and FOXA1 expression [109]. However, a couple of studies demonstrated the tumor-suppressive role of NEDD4 in colon cancer. For example, a clinical study discovered that NEDD4 is upregulated and NEDD4L downregulated in CRC tissues. Interestingly, patients with high expression of NEDD4L have shown to have longer disease-specific survival compared to patients with lower expression [76]. In CRC, NDRG1 prevented tumor growth by increasing p21 expression and decreasing its ubiquitylation. NDRG1 was found to be downregulated in CRC tissues, and a positive connection between NDRG1 and p21 was shown in vitro and in vivo. Mechanistically, NEDD4 directly interacts with p21 and targets it for degradation, which is inhibited by NDRG1, resulting in reduced growth of tumors. Further, it was demonstrated that the knockdown of NEDD4 with siRNA increased the expression of both p21 and NDRG1 [183]. Taken together, more studies are required for the proper understanding of the function of NEDD4 in CRC.
EC is the most frequent gynecologic malignancy in women, with an anticipated 61,880 new cases and 12,160 deaths in 2019 [223,224,225]. A study demonstrated the relationship between NEDD4 and EC, in which the cancer tissues showed increased levels of NEDD4. Moreover, immunohistochemistry analysis revealed strong NEDD4 staining, with higher NEDD4 expression in the most aggressive tumors. Furthermore, NEDD4 expression in EC was found to be favorably linked with the Akt downstream effector FoxM1 and increased cell growth, p-ERK, pAkt, and IGF-1R expression [155]. On the contrary, a clinical study showed considerably lower levels of NEDD4L expression in EC patients than those with benign endometrial diseases [154]. However, there is a paucity of evidence on the association between NEDD4 levels and the onset, progression, and prognosis of EC.
Gastric carcinoma (GC) is one of the most common and deadly cancers in the world [226,227,228]. A study revealed that NEDD4 was found to be expressed in GC and has tumorigenic activity by inducing ubiquitination and degradation of the PTEN [153]. To date, it has been reported that NEDD4 is overexpressed in 83 percentage of GC tumors. NEDD4 is probably involved in the signaling that leads to tumor metastasis caused by the EGFR because NEDD4 facilitates EGF-dependent GC cell invasion and migration [126]. Further, a high level of NEDD4L was associated with a low level of HIF-1α, indicating a positive prognosis [158]. In addition, another study revealed tumors that exhibited negative NEDD4L expression were highly linked to GC invasion and metastasis [157]. Another study demonstrated the prognostic value of this protein in GC. Decitabine (DAC), an inhibitor of DNA methylation, was found to increase NEDD4 expression in MGC803 GC cells, thus promoting the cells’ ability to invade and migrate [184]. However, to determine the role of NEDD4 in the development of GC, more in vitro, in vivo, and clinical research are needed.
In adults, glioblastoma is the most common form of adult brain cancer and is one of the top ten malignant tumors, despite surgery, radiation therapy, and chemotherapy, with an average nine-month survival time [161,229,230]. NEDD4 was shown to promote cell motility and invasion of malignant U251 glioma cells in vitro by ubiquitinating CNrasGEF [160]. Another study in U251 and U87 demonstrated that NEDD4 regulates glioma cell motility and invasion via the NEDD4/Rap2a pathway [161]. Moreover, it was shown that FOXM1 overexpression upregulated NEDD4, and the subsequent degradation of PTEN and activation of the Akt pathway in turn resulted in astrocyte transformation and GBM development [185]. Another study revealed that curcumin can reduce the expression of NEDD4, pAkt, and Notch1, which leads to the inhibition of cell growth and apoptosis and reduction in invasion and migration of glioma cells. Furthermore, deletion of NEDD4 sensitized glioma cells to curcumin [125]. Moreover, it was demonstrated that SMURF1 promotes the glioma cell migration and invasion and the expression of vimentin and MDM2, and the suppression of SMURF1 by siRNA transfection can reduce cell invasion and increase the p53, cleaved caspase-3, cleaved PARP, and E-cadherin [162]. Another study showed that miR-513a-5p repressed NEDD4L expression and participated in IGF-1 mediated activation of Wnt/β-catenin signaling, and consequently reduced TMZ cytotoxicity [186]. In human gliomas, the expression of NEDD4L is reduced, and patients with glioma who have low NEDD4L expression have a worse prognosis [159].
Liver cancer is an aggressive tumor that often develops from cirrhosis and chronic liver diseases [231,232]. Hepatocellular carcinoma (HCC), a primary liver cancer, is the third greatest cause of cancer-related death globally [233,234,235]. HCC, the most common type of liver cancer, is formed from hepatocytes and accounts for more than 80 percentage of all occurrences of liver cancer [236,237,238]. In HCC, the inhibition of cell proliferation, migration, and invasion, as well as cell cycle arrest in the S phase, was achieved by targeting NEDD4. Moreover, NEDD4 silencing resulted in an increase in PTEN expression, which resulted in decreased cell proliferation, migration, invasion, p-STAT3, p-Akt, and p-ERK1/2 [110]. In addition, it was observed that knockdown of NEDD4 decreased cell proliferation, viability, invasion, migration, and p-Akt levels and induced apoptosis and LATS1 expression in HCC [106]. In another study, it was reported that depletion or inhibition of NEDD4 could be used as a strategy to minimize metastasis and postpone tumor recurrence, hence improving survival rates. Moreover, NEDD4 knockdown in liver cancer cells increased PTEN and E-cadherin and decreased vimentin and p-Akt [111]. However, a study showed that NEDD4L suppressed cell growth by phosphorylating ERK1/2 and inducing apoptosis, suggesting that it might play a tumor-suppressive role in HCC by triggering MAPK/ERK-mediated apoptosis. Further, an in vivo study demonstrated that overexpression of NEDD4L prevented the growth of xenograft tumors in mice [163]. In another study, NEDD4 overexpression in HepG2 cells showed decreased guanylyl cyclase domain containing 1 (GUCD1) levels. Additionally, NEDD4 seems to influence GUCD1 degradation via the ubiquitin–proteasome system [187]. Therefore, further studies are required to confine the precise role of NEDD4 and to develop agonists/antagonists for the potential inhibition of HCC.
Lung cancer is the world’s top cause of death due to cancer in both men and women. It is a common cancer that is extremely aggressive and spreads quickly [14,239,240,241,242,243,244]. NSCLC accounts for 85 percentage of all lung cancers, with lung adenocarcinoma (ADC) accounting for 60 percentage of NSCLC, making ADC the most common histologic type [121,245,246,247]. It was reported that NEDD4 is linked with the progression of lung cancer and is overexpressed in a considerable proportion of NSCLC patients concerning chemosensitivity and prognosis. Thus, it promotes cell proliferation, migration, and invasion through PTEN degradation and increases the malignant characteristics of lung epithelial cells. Through stimulation of the lysosomal cathepsin B secretion pathway, NEDD4 can also mediate EGFR cell migration signaling in lung cancer cell lines [97,121,122]. A study in HCC827/ER cells demonstrated that NEDD4 might increase the acquired resistance of NSCLC cells to erlotinib by reducing the expression of PTEN. Moreover, knockdown of NEDD4 decreased tumor growth and tumor weight in nude mice (HCC827/ER cells) xenografts [146]. Another in vitro study showed that NEDD4L can act as a tumor suppressor, and it is downregulated in NSCLCs. Moreover, it was noticed that the knockdown of NEDD4L increased cell proliferation, invasion, and migration. Additionally, an in vivo study showed that knockdown of NEDD4L resulted in increased tumor growth and metastasis in nude mice (HCC827 LUC) [164]. Furthermore, another study demonstrated miR-93’s oncogenic role in lung cancer by downregulating the expression of NEDD4L. Reduced NEDD4L prevented SMAD2/SMAD3 from degrading, enhanced the TGF-β signal transduction, and promoted TGF-β-induced EMT [139]. Another study demonstrated that transmembrane prostate androgen-induced protein (TMEPAI) interacts with NEDD4 and binds to the TGFβ-type I receptor (TβRI), thereby facilitating its degradation. This study also showed that NEDD4 is necessary for the transport of TMEPAI to the lysosome [188]. NEDD4 is a particular E3 ligase for GCN2, which is activated in A549 cancer cells to increase tumor aggressiveness and survival for ubiquitination and degradation. The β-arrestin promotes the formation of the GCN2-β-arrestin-NEDD4L complex by binding NEDD4L to GCN2, enabling ubiquitin-mediated GCN2 degradation and thereby preventing cancer [189].
Human melanomas are the most aggressive form of malignant skin tumors, arising from neuro-ectodermal melanocytes. They can be astonishingly resistant to standard cancer therapy [191,248]. In line with this, a study showed that indole-3-carbinol (I3C) disruption of NEDD4 ubiquitination activity caused the wild-type PTEN tumor suppressor to stabilize and induced an antiproliferative response in melanoma [191,249]. An in vitro study proved that many melanoma cells show immunoreactivity for NEDD4L, and during the neoplastic transformation of melanocytes, the expression of NEDD4L might rise. Further, it was found that in xenograft models, exogenous NEDD4L expression significantly aided the growth of G-361 melanoma cells in vivo, and it was suggested that NEDD4L expression might be elevated in many melanomas to aid tumor growth [165]. Another study demonstrated that NEDD4 ubiquitinates immune checkpoint GITR and can promote melanoma cell proliferation by suppressing anti-tumor immune responses mediated by T cells [192]. A melanocytic transmembrane protein called Melan-A/MART-1 interacts with NEDD4 and ITCH and thus becomes ubiquitylated in melanoma cells. Both NEDD4 and ITCH contribute to the degeneration of the melanocyte. In pigmented cells, mutant Melan-A, which lacks ubiquitin acceptor residues, accumulates in melanosomes and has a prolonged half-life. Therefore, ubiquitylation regulates Melan-A/MART-1 lysosomal sorting and destruction from melanosomes [190]. In another study, a functional polypeptide JP1 was demonstrated to activate p-MEK1/2 and induced SP1 ubiquitination through the NEDD4L-SP1-Integrin αvβ3 pathway, which inhibited melanoma cell proliferation and metastases [193]. In A375 melanoma cell lines, erastin, a ferroptosis activator, stimulates the expression of NEDD4 and FOXM1 that leads to the ubiquitination and degradation of voltage-dependent anion channels, VDAC2/3, which further leads to the suppression of erastin-induced ferroptosis in melanoma. Additionally, an in vivo study showed that NEDD4 knockdown decreased tumor size and GSH levels, and increased MDA levels in nude mice (A375) xenografts [145]. Another study reported that NEDD4 promotes K48- and K63-dependent polyubiquitination and promotes lysosomal-dependent degradation of IGPR-1, because treatment of cells with lysosomal inhibitors such as bafilomycine enhanced IGPR-1 in human skin melanoma cell lines [194].
NEDD4 has been associated with the progression and development of nasopharyngeal carcinoma (NPC) [137,250]. In NPC cells, NEDD4 has tumorigenic capabilities because it promotes the EMT characteristics of cis-diamminedichloroplatinum (DDP, cisplatin) resistant cells. Overexpression of NEDD4 has been linked to EMT in DDP-resistant cells. Moreover, depletion of NEDD4 resulted in a partial reversion of EMT to MET phenotypes in resistant cells. These studies suggest that NEDD4 participates in EMT and chemoresistance in NPC cancer cells [137].
Neuroblastoma is the most frequently occurring extracranial solid tumor in children, and it is defined by the neoplastic development of neural crest cells in the developing sympathetic nervous system. The original tumor can occur anywhere along the sympathetic chain, although the adrenal gland is the most common site [251,252,253]. A recent study identified NEDD4 as a novel host component required for Japanese encephalitis virus (JEV) infection of human neural cells, and it was discovered that NEDD4 enhances JEV replication by reducing autophagy. Therefore, the downregulation of NEDD4 can drastically lower JEV infectivity, allowing neuronal cells to survive [195]. A study on neuroblastoma revealed that exosomal hsa-miR199a-3p overexpression in NB can increase proliferation and migration in vitro by inhibiting NEDD4 expression, resulting in a poor prognosis [196]. NEDD4 directly binds to Myc oncoproteins and targets them for ubiquitination and degradation. Additionally, small-molecule SIRT2 inhibitors activated the NEDD4 gene, decreased the expression of N-Myc and c-Myc proteins, and inhibited the proliferation of neuroblastoma cancer cells [103].
Dysregulation of E3 ubiquitin ligases appears to be a major component in the development and maintenance of ovarian cancer chemoresistance [254,255]. In ovarian cancer tissues, NEDD4L protein expression is found to be lower than non-cancer tissues. This suggests that the downregulation of NEDD4L protein expression might contribute to the emergence of ovarian cancer [166]. For example, DNA damage-binding protein 2 (DDBP2) modulates the responsiveness of ovarian cancer cells to TGF-β-induced growth suppression through NEDD4L [197]. Another in vitro study in OVCAR3 cells showed that erastin or RSL3 dose-dependently induced NEDD4L expression. Moreover, NEDD4L knockdown sensitized ovarian cancer cells to erastin- or RSL3-induced cell death and tumor suppression [198].
Pancreatic adenocarcinoma is an aggressive cancer of the pancreas that has a poor prognosis. PCa is the 14th most common cancer in the world and the 7th greatest cause of cancer death, with around 227,000 deaths each year expected around the world [256,257,258,259,260,261]. A growing number of studies have reported that NEDD4 has been linked to the development of PCa [262]. A surfeit of studies have proven that curcumin can exhibit its anti-tumor activity in cancer cells [263]. In PCa, curcumin downregulates the NEDD4 protein and further upregulates PTEN and p73, and led to the suppression of cancer. Importantly, it was noted that suppressing NEDD4 expression resulted in decreased cell proliferation, migration, and tumor growth [120]. Regarding LTF, an iron-binding transport protein, degradation by NEDD4L prevents intracellular iron accumulation and subsequent oxidative-damage-mediated ferroptosis cell death in PANC-1 cancer cells. In addition, this study showed that erastin or RSL3 dose-dependently enhanced NEDD4L expression, and NEDD4L knockdown effectively induced cell death and erastin or RSL3 stimulated MDA production [198]. Another study suggested that by directly interacting with the NEDD4 gene core promoter and deacetylating histone H4 lysine 16, SIRT2 inhibited the expression of the NEDD4 gene. Myc oncoproteins are directly targeted by NEDD4 for ubiquitination and degradation, and it was found that SIRT2 inhibitors activate the NEDD4 gene, which results in the decreased expression of Myc proteins, and decreased the PCa cell proliferation [103]. In conclusion, NEDD4 is a possible target for the therapy of PCa.
Prostate cancer is the second most observed cancer type in men and causes a series of health issues [202,264,265,266,267,268]. In the prostate gland, androgens and the androgen receptors (AR) play crucial roles for proper growth, differentiation, and physiological function. It has also been proven that AR dysregulation contributes to the advancement of prostate cancer [269,270,271,272]. In a clinical study, it was reported that the loss of NEDD4L expression was frequently observed in prostate cancer patients with more aggressive tumors. It was also shown that downregulation of NEDD4L significantly correlates with increased Gleason score, indicating the possibility of its role in prostate cancer progression [167]. A clinical study revealed that the expression of proteasomal pathway genes including PSMC4, PSMB5, and NEDD4L showed a significant upregulation. Interestingly, another member of this family, SMURF2, appears to be downregulated in organ-confined prostate tumors compared to non-organ-confined prostate tumors. Thus, dysregulation in the expression of these genes suggests a potential role in the development and spread of prostate cancer [168]. In another study, WWP2 was identified as an oncogene in prostate cancer. WWP2 influences PTEN’s degradation through a ubiquitylation-dependent pathway, regulating cellular apoptosis and increasing the development of cancer in a nude mice (DU145) xenograft, and the knockdown of WWP2 was shown to decrease tumor growth [173,199]. A negative link was discovered between HER3 and NEDD4 levels in the DU145 prostate cancer cell lines. More importantly, NEDD4 knockdown enhanced HER3 expression, making cancer cells sensitive to an anti-HER3 antibody which limited cell proliferation [172]. In conclusion, NEDD4 could be a possible target for the therapy of prostate cancer.
Apart from these studies, the role of NEDD4 has also been studied in other cancers, such as multiple myeloma (MM), uveal melanoma, bone cancer, gall bladder cancer, etc. [156,169,273,274]. For example, NEDD4 has also been involved in the development of gallbladder carcinoma, which is an aggressive cancer with a high rate of invasiveness that continues to have a poor prognosis, with a less than 10 percentage overall survival rate [156]. A study highlighted NEDD4 as an invasion-associated molecule in gallbladder carcinoma and demonstrated that at the transcriptional level NEDD4L regulates MMP-1 (Matrix metalloproteinase-1) and MMP-13 expression, which is overexpressed at an early stage of tumor invasion in various malignant tumors [156]. Further, a study in U2OS cells reported that NEDD4L promotes the ubiquitylation of 8-Oxoguanine DNA glycosylase (OGG1), a major cellular enzyme involved in the base excision repair pathway, which plays a vital role in suppressing mutagenesis and controls genome stability [169].
A surfeit of studies has found that NEDD4 exhibits a potential oncogenic role and is frequently overexpressed in various types of cancers. For example, in an in vivo study, it was reported that overexpression of NEDD4 enhanced the growth and metastasis of xenograft tumors, where it was abundantly expressed in colon cancer tissues and cells [109]. Similarly, the expression of NEDD4 is higher in malignant endometrium compared to benign. The levels of FOXM1, an oncogenic transcription factor, are positively correlated with NEDD4 expression in endometrial cancers. In addition, NEDD4 overexpression in endometrial cancer cells enhanced IGF-1R cell surface localization, Akt activation, and cell proliferation [155]. In GCA, NEDD4 is overexpressed, and this is linked to tumor invasion and metastasis, as well as survival rate in rats [126]. Moreover, overexpression of NEDD4 enhanced cell proliferation, reduced cell apoptosis, and promoted cell invasion and migration in bladder cancer cells. In addition, it lowered PTEN levels while increasing Notch-1 expression [105]. NEDD4 is a factor that mediates PTEN ubiquitination and inactivation, and in human NSCLC, overexpression of this protein causes a rise in the ubiquitylated form of PTEN, lowering PTEN levels and enhancing Akt activation [146]. Moreover, another study demonstrated that NEDD4 overexpression was associated with poor clinical outcomes in patients with HCC and that NEDD4 depletion inhibited proliferation, migration, and invasion of Huh7 cells through the upregulation of PTEN [111]. Interestingly, it was found that anti-tumor immunity mediated by T cells against melanoma cells is inhibited when NEDD4 is overexpressed. Moreover, the expression of NEDD4 was shown to be higher in metastatic melanoma tissues and is linked with bad prognosis [192]. Thus, these data suggest that targeting NEDD4 for cancer treatment could be a promising method.
NEDD4 also exerts tumor suppressor properties in various malignancies. For example, NEDD4L prevents autophagy activity under metabolic stress by destabilizing ULK1, an autophagy protein, and ASCT2, a glutamine transporter, hence reducing mitochondrial functioning and tumor suppression [75]. Similarly, shRNA inactivation of NEDD4 increases HER3 levels in prostate cancer cells, increasing HER3 signaling and cancer cell growth [172]. Another study revealed that NEDD4 lowered the Myc protein synthesis and decreased proliferation of neuroblastoma and pancreatic cancer cells by directly binding to Myc oncoproteins and targeting them for ubiquitination and degradation [103]. Additionally, NEDD4 deregulated the PIP5Kα-dependent PIP2 pool that promotes proliferation in BC cells via PI3K/Akt activation [113].
Several investigations in knockout mouse models have shown the biological activities of NEDD4. For example, an in vivo study showed decreased IGF-1 and insulin signaling, delayed embryonic development, reduced growth and body weight, and neonatal mortality in NEDD4 null mice. In addition, homozygous NEDD4 knockout in mice reduced mitogenic activity in embryonic fibroblasts and increased the amount of the adaptor protein Grb10, and mislocalized the IGF-1 receptor that is usually found on the plasma membrane. Thus, NEDD4 appears to favorably modulate IGF-1 and insulin signaling in vivo, in part through regulating Grb10 activity [275]. Consistent with this report, another study on NEDD4 null mutant mice delineated that NEDD4 deficiency results in perinatal mortality and aberrant neuromuscular structure and function. This showed that NEDD4 is essential for survival throughout mammalian embryonic development [276]. Another study showed that NEDD4 knockout in mice leads to embryonic mortality at mid-gestation, with severe heart malformations and vascular abnormalities. Further, elevated levels of thrombospondin-1 (Tsp-1), an angiogenesis inhibitor, were seen in knockout mouse embryonic fibroblasts and embryos. Interestingly, administration of aspirin (a Tsp-1 inhibitor) to pregnant heterozygote mothers resulted in lower Tsp-1 levels and a significant reduction in embryonic lethality. These findings show that NEDD4 is a Tsp-1 suppressor and that elevated Tsp-1 levels in NEDD4 knockout mice might have contributed to the developmental abnormality seen in the embryos [277]. Another study reported that NEDD4 knockout causes dendrites to become shorter and less complex, resulting in a decreased number of functional synapses and, consequently, synaptic transmission [278]. In another study, the conditional deletion of NEDD4L in lung epithelial cells led to chronic lung disease, which has many of the characteristics of IPF, such as bronchiolization and progressive fibrosis, Muc5b overexpression in honeycombing, peripheral airways, and distinctive alterations in the lung proteome [96]. Therefore, these results indicate that complete deletion of NEDD4 is lethal. Although NEDD4 depletion does not promote tumor development on its own, it does cause tumor growth enhancement in Apc+/min-derived colorectal tumors, implying that NEDD4 typically reduces intestinal Wnt signaling and colonic tumorigenesis. These findings imply that NEDD4 inhibits colonic Wnt signaling and tumor growth, at least in part, by inhibiting the transcription factors LEF1 and YY1 [279]. The precise role of NEDD4 in tumorigenesis must be further validated.
In several studies, NEDD4 has been proven to promote tumor growth, and thus targeting this protein for cancer treatment is a promising method. For example, in glioma cells, curcumin, a diarylheptanoid ingredient found in Curcuma longa plants, decreased proliferation, invasion, and migration by suppressing NEDD4 [125]. In addition, treatment of indole-3-carbino (I3C), a naturally occurring compound found in cruciferous vegetables, along with knockdown of NEDD4, inhibited degradation of wild-type PTEN, causing antiproliferation in melanoma cells [191]. Moreover, diosgenin, a saponin produced from the Trigonellafoenum graecum plant, inhibited NEDD4 expression in prostate cancer cells, resulting in anticancer action [268]. Further, Decitabine, a DNA methylation inhibitor, increased migration and invasion of gastric cancer cells by increasing the expression of NEDD4 [184]. As a result, targeting NEDD4 could be an essential therapeutic approach for the management of human cancers.
By combining studies, it is clear that NEDD4/NEDD4-like E3 ligases are crucial to the multistep process that causes different forms of cancer. In most cancer types, NEDD4 inhibits tumorigenesis by increasing the degradation of NEDD4 substrates which have a crucial oncogenic role in various malignancies; however, its significance in a few cancer types is still debated. Therefore, due to its role in tumorigenesis, NEDD4 can be a potential target in drug discovery against various types of cancer. Because of NEDD4’s many substrates and dual functions, treatment approaches that disrupt NEDD4’s interactions with these substrates with no/fewer side effects might be more suitable than those that directly target NEDD4 activity. However, further studies are required to completely understand the relevance of NEDD4 in tumorigenesis. | true | true | true |
|
PMC9604297 | Silvia Romano,Carmela Romano,Martina Peconi,Alessia Fiore,Gianmarco Bellucci,Emanuele Morena,Fernanda Troili,Virginia Cipollini,Viviana Annibali,Simona Giglio,Rosella Mechelli,Michela Ferraldeschi,Liana Veneziano,Elide Mantuano,Gabriele Sani,Andrea Vecchione,Renato Umeton,Franco Giubilei,Marco Salvetti,Rosa Maria Corbo,Daniela Scarabino,Giovanni Ristori | Circulating U13 Small Nucleolar RNA as a Potential Biomarker in Huntington’s Disease: A Pilot Study | 18-10-2022 | fluid biomarkers,small circulating non-coding RNAs,small nucleolar RNAs,Huntington’s disease | Plasma small RNAs have been recently explored as biomarkers in Huntington’s disease (HD). We performed an exploratory study on nine HD patients, eight healthy subjects (HS), and five psychiatric patients (PP; to control for iatrogenic confounder effects) through an Affymetrix-Gene-Chip-miRNA-Array. We validated the results in an independent population of 23 HD, 15 pre-HD, 24 PP, 28 Alzheimer’s disease (AD) patients (to control the disease-specificity) and 22 HS through real-time PCR. The microarray results showed higher levels of U13 small nucleolar RNA (SNORD13) in HD patients than controls (fold change 1.54, p = 0.003 HD vs. HS, and 1.44, p = 0.0026 HD vs. PP). In the validation population, a significant increase emerged with respect to both pre-HD and the control groups (p < 0.0001). SNORD13 correlated with the status of the mutant huntingtin carrier (r = 0.73; p < 0.001) and the disease duration (r = 0.59; p = 0.003). The receiver operating characteristic (ROC) curve analysis showed the high accuracy of SNORD13 in discriminating HD patients from other groups (AUC = 0.963). An interactome and pathway analysis on SNORD13 revealed enrichments for factors relevant to HD pathogenesis. We report the unprecedented finding of a potential disease-specific role of SNORD13 in HD. It seems to peripherally report a ‘tipping point’ in the pathogenic cascade at the neuronal level. | Circulating U13 Small Nucleolar RNA as a Potential Biomarker in Huntington’s Disease: A Pilot Study
Plasma small RNAs have been recently explored as biomarkers in Huntington’s disease (HD). We performed an exploratory study on nine HD patients, eight healthy subjects (HS), and five psychiatric patients (PP; to control for iatrogenic confounder effects) through an Affymetrix-Gene-Chip-miRNA-Array. We validated the results in an independent population of 23 HD, 15 pre-HD, 24 PP, 28 Alzheimer’s disease (AD) patients (to control the disease-specificity) and 22 HS through real-time PCR. The microarray results showed higher levels of U13 small nucleolar RNA (SNORD13) in HD patients than controls (fold change 1.54, p = 0.003 HD vs. HS, and 1.44, p = 0.0026 HD vs. PP). In the validation population, a significant increase emerged with respect to both pre-HD and the control groups (p < 0.0001). SNORD13 correlated with the status of the mutant huntingtin carrier (r = 0.73; p < 0.001) and the disease duration (r = 0.59; p = 0.003). The receiver operating characteristic (ROC) curve analysis showed the high accuracy of SNORD13 in discriminating HD patients from other groups (AUC = 0.963). An interactome and pathway analysis on SNORD13 revealed enrichments for factors relevant to HD pathogenesis. We report the unprecedented finding of a potential disease-specific role of SNORD13 in HD. It seems to peripherally report a ‘tipping point’ in the pathogenic cascade at the neuronal level.
Huntington’s disease (HD) is an inherited neurodegenerative disease caused by CAG trinucleotide repeat expansion in the first exon of the HTT gene, which encodes the huntingtin protein. HD is a progressive, incurable disease with a typical adult onset, related to CAG repeat length, and is characterized by motor impairment, cognitive dysfunction, and psychiatric symptoms. Disease-modifying treatments for HD are under development, and the identification of easily measurable biomarkers is crucial for predicting disease progression, monitoring the effects of novel drugs, and obtaining cues on the pathogenic cascade at the neuronal level. Peripheral biomarkers are quantified in body fluids with minimal invasiveness, good accuracy, and a high discriminatory power. Cerebrospinal fluid (CSF) has been a focus of interest as a proxy for central nervous system (CNS) pathophysiology, and recent studies have identified reliable biomarkers, such as CSF mutant huntingtin (mHTT), used as an outcome measure for therapeutic approaches [1,2], which showed a very good predictive power for disease manifestation [3,4]. Biomarkers based on complex techniques or CSF may be of limited use because of their invasiveness, high cost, or the need for specialized personnel. In recent times, increased focus has been given to the testing of more easily measurable biomarkers from peripheral leukocytes and plasma; these are cheaper, less invasive, and potentially more adept in obtaining longitudinal profiles. Among these, the measurement of leukocyte telomere length (LTL), which shortens remarkably in pre-symptomatic HD (PreHD) is a possible measure of time to clinical onset [5]. The histone variant pγ-H2AX is a component of the DNA damage responses in peripheral blood mononuclear cells (PBMCs) and changes in its levels proved to be an informative, potentially reversible biomarker in pre-HD [6]. In other studies, the PBMCs from patients with HD have been used to study mHTT and other gene expression profiles as predictors of disease progression [7,8]. Plasma neurofilament light protein (NfL) is reliable in monitoring disease progression, although it is less sensitive than CSF NfL [4]. Other studies have reported informative results on the plasma levels of oxidative stress markers, metabolic markers, and immune system products [9]. Recent studies on small non-coding RNAs in the plasma of patients with HD have led to several investigations of circulating micro-RNAs (miRNAs) [10,11,12]. Our recent study found that hsa-miR-323b-3p is upregulated in individuals with an mHTT mutation [13]. In this context, we obtained results on small nucleolar RNAs (snoRNAs), which have not been previously studied in HD, and thus, we considered snoRNAs as a possible peripheral biomarker of disease and elucidated their role in disease pathogenesis and progression. SnoRNAs are a class of non-coding small guide RNAs, most of which direct the chemical modifications of other RNA substrates, including ribosomal RNAs (rRNAs) and spliceosomal RNAs. Moreover, some snoRNAs are involved in the regulation of alternative splicing and post-transcriptional modifications of mRNA [14]. Homo sapiens U13 snoRNA (SNORD13), 104 nucleotides long, is a member of the Box C/D family of small nucleolar ribonucleoproteins that can form base-pair interactions with the 3’ portion of 18S rRNA and is involved in the processing of this rRNA [15]. This study is aimed at reporting changes in circulating SNORD13 levels in people with prodromal and overt HD compared with several groups of controls in order to avoid confounders and to verify the disease specificity of our finding.
We performed an exploratory microarray study of whole noncoding RNA expression profiles in the plasma of nine patients with HD (mean age of 48.25 ± 10.47; four males and five females), and 13 controls including eight healthy subjects (HS, mean age of 49.17 ± 11.79; two males and six females) and five psychiatric patients (PP, mean age of 50.25 ± 11.47; two males and three females) with schizophrenia or bipolar disorder. As these are the patients with HD that are often treated with psychotropic drugs, we included PP with similar treatment profiles as a control group in order to minimize the possible iatrogenic impact on the profile of the small non-coding RNAs. In particular, we selected patients treated with olanzapine, lithium, and valproate which were the treatments most frequently prescribed to patients affected by HD. The microarray results indicated that SNORD13 levels were increased in the plasma of patients with HD compared to those in the HS and PP control groups (fold change, 1.54; p = 0.0003 HD vs. HS, and fold change, 1.44; p = 0.0026 HD vs. PP; Figure 1A). To validate this result, SNORD13 plasma levels were quantified using real-time PCR in five cohorts of subjects: 22 HS, 23 symptomatic patients with HD, 15 patients with pre-manifest HD (pre-HD), 24 PP, and 28 patients with Alzheimer’s disease (AD). The last group was considered as a control for the disease specificity of our findings. Demographic and clinical characteristics of each group are shown in Table 1. No significant relationship was observed between SNORD13 plasma levels and age/sex at blood sampling in any group, except for females in the HS group (p = 0.04). No relationship was observed between SNORD13 levels and CAG repeat length in the subjects with pre-HD or HD (Table 2). Our analysis showed a statistically significant (p < 0.0001) increase in the plasma levels of SNORD13 in patients with HD, which clearly segregated patients with overt disease (HD) from controls and pre-HD subjects. The changes in the plasma level of SNORD13 in symptomatic HD patients were highly significant compared to those of both the pre-HD and the three control groups (HS, PP, and AD; Figure 1B, Table 2). A positive linear correlation was observed between circulating SNORD13 levels and disease duration in patients with HD (r = 0.589, p = 0.003) (Figure 2A). A significant relationship was also observed between plasma SNORD13 and the UHDRS clinical score in mHTT carriers (r = 0.732, p < 0.001, Figure 2B), whereas it was not found to be significant in only overt patients with HD (not shown). These linear relationships remained significant when considering age as a covariate (partial correlations, r = 0.563, p = 0.01, and r = 0.685, p < 0.001, respectively). Next, we assessed the accuracy of plasma SNORD13 as a biomarker of overt HD through ROC curve analysis. In discriminating symptomatic HD patients from pre-symptomatic HTT mutation carriers, SNORD13 displayed an extremely high accuracy (AUC = 0.963, Figure 3A); setting the cut-off point of SNORD13 levels at 0.58 allowed the identification of patients with HD with 95.88% sensitivity and 86.7% specificity. Moreover, SNORD13 appeared to be of potential utility in distinguishing symptomatic HD patients from control groups (AD, PP, HS; AUC = 0.953; Figure 3B), as well as pre-HD among controls (AUC = 0.955; Figure 3C), and to a lesser extent, in identifying HTT mutation carriers (AUC: 0.811; Figure 3D). Finally, to investigate the biological landscape of the action of SNORD13, we constructed an interactome. We retrieved information on U13 snoRNA interactions with proteins (including known RNA-binding proteins (RBPs) and transcription factors (TF)), other snoRNAs, miRNAs, and rRNAs from the databases RNAinter [16], snoDB [17], and RNAct [18]. Additionally, we mapped the intra-network protein-protein interactions through STRING [19]. The final network comprised 456 SNORD3-interacting nodes: 258 TFs, 86 RBPs, 91 proteins, one long noncoding RNA, three miRNAs, 13 snoRNAs, and the 18 s rRNA ribosomal subunit (Figure 4A and Supplementary Table S1). A pathway analysis revealed enrichment of processes involved in transcriptional regulation and RNA metabolism (Figure 4B–D), referring to molecules mostly located in the nucleus and involved in genomic organization (Figure 4E). Of interest in HD is the emergence of nerve growth factor (NGF)-stimulated transcription associated with SNORD13 activity, suggesting a direct implication in neurodegenerative processes, as well as the interaction with molecules involved in the DNA damage response that has already been implicated in disease pathogenesis and that are useful as peripheral biomarkers [6,7,20].
Our study highlights an unprecedented finding of the potential role of snoRNAs in HD. The main results are the following: (i) the levels of circulating SNORD13 are significantly increased in the overt disease compared to the prodromal phase of HD; (ii) the levels of this snoRNA are comparable between HS and patients with a pre-HD status; (iii) this finding seems specific for HD, since three groups of control (HS, PP on drugs similar to that administered to patients with HD, and patients with AD) showed normal, comparable values; (iv) the plasma levels of SNORD13 seems to be related to the natural history of HD, correlating with the status of mHTT carriers and the disease duration; and (v) the above points (iii) and (iv) suggest that increased levels of SNORD13 in blood mirror pathogenic events in the CNS (though this fact awaits formal demonstration), possibly paving the way for new therapeutic targets. A possible explanation for our data is that the elevated plasma level of SNORD13 in symptomatic HD patients may be due to nucleolar stress caused by the presence of mutant RNAs that carry an expanded CAG repeat (expanded CAG RNAs). In recent years, increasing attention has been directed toward understanding the pathogenic mechanisms of mHTT, and several studies have supported the hypothesis that expanded CAG RNAs induce apoptosis by activating the nucleolar stress pathway in both patients with HD and transgenic models of the disease [21,22]. Specifically, it has been shown that expanded CAG RNAs compete with nucleolin (a multifunctional protein that is mainly localized in the nucleolus and involved in various steps of ribosome biogenesis) for the rRNA promoter, leading to the reduction of rRNA expression, nucleolar stress and apoptosis via p53, and activation of the downstream signaling cascade, including mitochondrial cytochrome C release and caspase activation [23,24,25,26]. However, considering the SNORD13 interactome, we cannot rule out the possibility that other pathogenic mechanisms might be involved in the change in the levels of snoRNAs. Indeed, beyond ribosomal biogenesis and RNA metabolism, SNORD13 appears to be involved in a wide range of genomic activities associated with HD pathophysiology (Figure 4, Supplementary Table S1): chromatin remodeling via histone acetylation and methylation (SIRT6, EP300, TAF1, CHD1, KDM5A-B etc.), telomere length maintenance (DKC1,HRNPU, VPRBP) [5], DNA repair and damage response (ERCC6, BAZ1B, FANCD2, TOPBP1) [6,20]; direct modulation of homeobox (HOXA1-7, DLX1-2, OTX2 and others), and zinc-finger proteins (SNAI2, SP1-2, ZMIZ2 and others) [27]. We also identified three miRNAs (hsa-miR-455-5p, hsa-miR-342-3p, and hsa-miR-377-3p) which may cooperate with U13 snoRNA in regulating gene expression as being possibly affected in HD; stochastic computational analyses might help further explain these relationships, as piloted in multiple sclerosis pathogenesis modeling [28], COVID epidemic wave stability evaluation [29], and many other complex problems. Irrespective of the mechanism(s) possibly linking mHTT to SNORD13 and the source of this snoRNA (currently unknown), we can speculate that the increase in the plasma level of SNORD13 in patients with HD may peripherally report a ‘tipping point’ in the pathogenic cascade at the neuronal level, while normal levels may mark a ‘molecular pre-manifest status’ in disease evolution. In fact, when we used published data [2] on CSF mHTT and plasma or CSF NfL as benchmarks to compare the performance of plasma SNORD13, we found that it outperformed both CSF mHTT and plasma or intrathecal NfL as a reporter of overt HD, supporting its potential value as a peripheral cue of central pathogenic processes (Table 3) In a clinical context circulating SNORD13 does not have a predictive value, being absent in pre-HD; however, this easily measurable, inexpensive and reliably quantifiable test (it is consistent, accurate, sensitive, specific, and reproducible) [9] may be useful to better catch, on a laboratory basis, the shift from prodromal to overt HD [30]. This may be useful to stratify and select candidate patients for clinical trials and to aim at disease-specific pathophysiological cascades. In summary, circulating SNORD13 could be a clinically actionable asset in future trials as a peripheral reporter of a ‘functional reserve’ at the neuronal level, as well as a hint for new therapeutic targets. Our research is subject to several limitations: the first is the small sample size, but we performed a sample size calculation to reduce this type of bias; the second limitation is the selection bias due to the different age of onset of each disease group taken into exam (AD patients are usually older than HD patients while PP are usually younger); and lastly, our results derive from a cross-sectional study and they consequentially reflect SNORD13 plasma levels at a single point in time both in HD, pre-HD and controls. Further longitudinal studies with a larger sample size will be necessary to reduce these limitations and to validate our hypothesis. In conclusion, peripheral small RNAs may offer potential advantages in terms of disease specificity compared to other approaches, such as circulating NfL, which has already been reported as a promising peripheral biomarker for many neurological diseases [3]. Future personalized therapies for different phases of the HD course and possible trials with etiologic approaches in subjects with pre-HD are necessary, which will plausibly introduce circulating biomarkers such as SNORD13 as promising components of the neurogeneticist’s toolkit.
Participants in the study were enrolled at the Center for Experimental Neurological Therapies, Unit of Neurology (patients with a positive test for HD, patients with probable Alzheimer’s disease, and healthy subjects), and the School of Medicine and Psychology, Unit of Psychiatry (patients with psychiatric disorders), S. Andrea Hospital, Department of Neurosciences, Mental Health, and Sensory Organs, Sapienza University of Rome, Italy. The study was approved by the local ethics committee and written consent was obtained from all participants according to the principles of the Declaration of Helsinki. The operators were unaware of the disease state of each sample during processing and statistical analysis. The eligible subjects for this study were patients with a positive genetic test for HD, a diagnosis of schizophrenia or bipolar disorders treated with antipsychotic drugs, a diagnosis of probable Alzheimer’s disease, and healthy subjects (controls). Exclusion criteria were pregnancy, breastfeeding, and severe systemic illnesses or conditions.
Blood samples were obtained via venous puncture in ethylenediaminetetraacetic acid (EDTA) tubes for plasma preparation. The plasma was obtained by centrifugation (1500× g for 15 min at 4 °C); a few 500 μL aliquots of supernatant were stored at −80 °C. RNA was extracted using a Plasma/Serum Circulating RNA Purification Kit (NORGEN) following the manufacturer’s instructions. RNA quality and purity were assessed using an RNA 6000 Nano Assay kit on an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Briefly, 500 ng of total RNA was labeled using FlashTag Biotin HSR (Genisphere, Hatfield, PA, USA) and hybridized to GeneChipR miRNA 2.0 Arrays. The arrays were stained in Fluidics Station 450 and scanned using a GeneChip R Scanner 3000 (Affymetrix, Santa Clara, CA, USA). A statistical analysis was performed using Transcriptome Analysis Console (TAC) software (Thermo Fisher Scientific, Waltham, MA, USA). To survey the presence of outliers that could impact the dataset, principal component analysis (implemented in R) was performed to identify possible outliers that needed to be excluded. MiRNA probe outliers were defined according to the manufacturer’s instructions (Affymetrix), and further analysis included data summarization, normalization, and quality control using the web-based miRNA QC Tool software (Affymetrix). The raw dataset is available from the Gene Expression Omnibus (GEO) repository (GSE167630).
RNA extraction was performed from plasma using the miRNeasy Serum/Plasma kit according to the manufacturer’s protocol (Qiagen, Hilden, Germany). A U13 snoRNA analysis was performed by quantitative reverse transcription PCR (qRTPCR). cDNA was synthesized using the TaqMan™ MicroRNA Reverse Transcription Kit (Life Technologies, Carlsbad, CA, USA), according to the manufacturer’s instructions using a U13 snoRNA-specific primer (manufacturer-provided). Quantitative real-time PCR was performed on an ABI 7300 Real-time PCR System (using custom TaqMan® Small RNA Assay (Life Technologies), according to the manufacturer’s instructions. Relative quantitation of U13 snoRNA was performed by the delta Ct method, using U6 snoRNA as an endogenous control [32]. Replicate assays of the same sample were performed to calculate inter-assay variation. The average standard deviation (SD) calculated by measuring the plasma SNORD13 levels of a sample repeated over three different assays was 0.035%. Thus, assuming a normal distribution, samples differing in average SNORD13 levels by as little as 0.069% (1.96 × SD) should be distinguishable by this method at a 95% confidence interval [33].
Statistical analyses were performed using Partek Genomic Suite software (miRNA Array data), GraphPad Prism v9, and R (The R Project for Statistical Computing) v3.6.3. Data normality was assessed using the Shapiro–Wilk test. Nonparametric tests were used to compare the distribution of SNORD13 plasma levels between the patients and controls. Nonparametric tests or linear regression were used to evaluate the distribution of SNORD13 plasma levels across age, sex, and CAG repeat number. Spearman’s correlation was computed to assess the linear relationship between SNORD13 plasma levels and disease duration or the Unified Huntington’s Disease Rating Scale-Total Motor Score (UHDRS-TMS), and age was considered a covariate in the partial correlations. Statistical significance was set at p < 0.05. The sample size calculation was based on our preliminary study that enrolled 10 HD subjects and 10 controls, detecting a difference between group means of 0.58 with a standard deviation of 0.34. Assuming α = 0.01 and equal group size, we would need to study 12 HD and 12 control subjects to reject the null hypothesis that the population means of the experimental and control groups are equal with probability (power) 0.9. A receiver operating characteristic (ROC) curve analysis was performed using the R easy ROC web interface. The optimal cutoff point was identified according to the Youden index method. Sample size estimation settings were as follows: type I error, 0.05; power, 0.8; allocation ratio, 1; area under the curve (AUC) derived from the HD/pre-HD ROC curve.
The molecular interactions of SNORD13 were screened in three databases: (1) the RNA interactome database (RNAinter v4.0) [34] at http://www.rna-society.org/rnainter/ (accessed on 1 November 2021), which includes more than 41 million predicted or experimentally validated interactions; (2) snoDB, the largest repository of snoRNA biological annotation and manually curated snoRNA-RNA interactions [35]; (3) RNAct, a database of genome-wide predicted protein–RNAs interactions [17]. Nonhuman interactors were excluded from this study. Protein-protein interactions were derived by querying the STRING database [18] for high-confidence interactions (score > 0.7) of SNORD13-interacting proteins. The global SNORD13 interactome was mapped using Cytoscape [19]. A pathway enrichment analysis was performed using the STRING application in Cytoscape. Graphical plotting was performed using the ggplot2 package in R [20]. | true | true | false |
|
PMC9604297 | Silvia Romano,Carmela Romano,Martina Peconi,Alessia Fiore,Gianmarco Bellucci,Emanuele Morena,Fernanda Troili,Virginia Cipollini,Viviana Annibali,Simona Giglio,Rosella Mechelli,Michela Ferraldeschi,Liana Veneziano,Elide Mantuano,Gabriele Sani,Andrea Vecchione,Renato Umeton,Franco Giubilei,Marco Salvetti,Rosa Maria Corbo,Daniela Scarabino,Giovanni Ristori | Circulating U13 Small Nucleolar RNA as a Potential Biomarker in Huntington’s Disease: A Pilot Study | 18-10-2022 | fluid biomarkers,small circulating non-coding RNAs,small nucleolar RNAs,Huntington’s disease | Plasma small RNAs have been recently explored as biomarkers in Huntington’s disease (HD). We performed an exploratory study on nine HD patients, eight healthy subjects (HS), and five psychiatric patients (PP; to control for iatrogenic confounder effects) through an Affymetrix-Gene-Chip-miRNA-Array. We validated the results in an independent population of 23 HD, 15 pre-HD, 24 PP, 28 Alzheimer’s disease (AD) patients (to control the disease-specificity) and 22 HS through real-time PCR. The microarray results showed higher levels of U13 small nucleolar RNA (SNORD13) in HD patients than controls (fold change 1.54, p = 0.003 HD vs. HS, and 1.44, p = 0.0026 HD vs. PP). In the validation population, a significant increase emerged with respect to both pre-HD and the control groups (p < 0.0001). SNORD13 correlated with the status of the mutant huntingtin carrier (r = 0.73; p < 0.001) and the disease duration (r = 0.59; p = 0.003). The receiver operating characteristic (ROC) curve analysis showed the high accuracy of SNORD13 in discriminating HD patients from other groups (AUC = 0.963). An interactome and pathway analysis on SNORD13 revealed enrichments for factors relevant to HD pathogenesis. We report the unprecedented finding of a potential disease-specific role of SNORD13 in HD. It seems to peripherally report a ‘tipping point’ in the pathogenic cascade at the neuronal level. | Circulating U13 Small Nucleolar RNA as a Potential Biomarker in Huntington’s Disease: A Pilot Study
Plasma small RNAs have been recently explored as biomarkers in Huntington’s disease (HD). We performed an exploratory study on nine HD patients, eight healthy subjects (HS), and five psychiatric patients (PP; to control for iatrogenic confounder effects) through an Affymetrix-Gene-Chip-miRNA-Array. We validated the results in an independent population of 23 HD, 15 pre-HD, 24 PP, 28 Alzheimer’s disease (AD) patients (to control the disease-specificity) and 22 HS through real-time PCR. The microarray results showed higher levels of U13 small nucleolar RNA (SNORD13) in HD patients than controls (fold change 1.54, p = 0.003 HD vs. HS, and 1.44, p = 0.0026 HD vs. PP). In the validation population, a significant increase emerged with respect to both pre-HD and the control groups (p < 0.0001). SNORD13 correlated with the status of the mutant huntingtin carrier (r = 0.73; p < 0.001) and the disease duration (r = 0.59; p = 0.003). The receiver operating characteristic (ROC) curve analysis showed the high accuracy of SNORD13 in discriminating HD patients from other groups (AUC = 0.963). An interactome and pathway analysis on SNORD13 revealed enrichments for factors relevant to HD pathogenesis. We report the unprecedented finding of a potential disease-specific role of SNORD13 in HD. It seems to peripherally report a ‘tipping point’ in the pathogenic cascade at the neuronal level.
Huntington’s disease (HD) is an inherited neurodegenerative disease caused by CAG trinucleotide repeat expansion in the first exon of the HTT gene, which encodes the huntingtin protein. HD is a progressive, incurable disease with a typical adult onset, related to CAG repeat length, and is characterized by motor impairment, cognitive dysfunction, and psychiatric symptoms. Disease-modifying treatments for HD are under development, and the identification of easily measurable biomarkers is crucial for predicting disease progression, monitoring the effects of novel drugs, and obtaining cues on the pathogenic cascade at the neuronal level. Peripheral biomarkers are quantified in body fluids with minimal invasiveness, good accuracy, and a high discriminatory power. Cerebrospinal fluid (CSF) has been a focus of interest as a proxy for central nervous system (CNS) pathophysiology, and recent studies have identified reliable biomarkers, such as CSF mutant huntingtin (mHTT), used as an outcome measure for therapeutic approaches [1,2], which showed a very good predictive power for disease manifestation [3,4]. Biomarkers based on complex techniques or CSF may be of limited use because of their invasiveness, high cost, or the need for specialized personnel. In recent times, increased focus has been given to the testing of more easily measurable biomarkers from peripheral leukocytes and plasma; these are cheaper, less invasive, and potentially more adept in obtaining longitudinal profiles. Among these, the measurement of leukocyte telomere length (LTL), which shortens remarkably in pre-symptomatic HD (PreHD) is a possible measure of time to clinical onset [5]. The histone variant pγ-H2AX is a component of the DNA damage responses in peripheral blood mononuclear cells (PBMCs) and changes in its levels proved to be an informative, potentially reversible biomarker in pre-HD [6]. In other studies, the PBMCs from patients with HD have been used to study mHTT and other gene expression profiles as predictors of disease progression [7,8]. Plasma neurofilament light protein (NfL) is reliable in monitoring disease progression, although it is less sensitive than CSF NfL [4]. Other studies have reported informative results on the plasma levels of oxidative stress markers, metabolic markers, and immune system products [9]. Recent studies on small non-coding RNAs in the plasma of patients with HD have led to several investigations of circulating micro-RNAs (miRNAs) [10,11,12]. Our recent study found that hsa-miR-323b-3p is upregulated in individuals with an mHTT mutation [13]. In this context, we obtained results on small nucleolar RNAs (snoRNAs), which have not been previously studied in HD, and thus, we considered snoRNAs as a possible peripheral biomarker of disease and elucidated their role in disease pathogenesis and progression. SnoRNAs are a class of non-coding small guide RNAs, most of which direct the chemical modifications of other RNA substrates, including ribosomal RNAs (rRNAs) and spliceosomal RNAs. Moreover, some snoRNAs are involved in the regulation of alternative splicing and post-transcriptional modifications of mRNA [14]. Homo sapiens U13 snoRNA (SNORD13), 104 nucleotides long, is a member of the Box C/D family of small nucleolar ribonucleoproteins that can form base-pair interactions with the 3’ portion of 18S rRNA and is involved in the processing of this rRNA [15]. This study is aimed at reporting changes in circulating SNORD13 levels in people with prodromal and overt HD compared with several groups of controls in order to avoid confounders and to verify the disease specificity of our finding.
We performed an exploratory microarray study of whole noncoding RNA expression profiles in the plasma of nine patients with HD (mean age of 48.25 ± 10.47; four males and five females), and 13 controls including eight healthy subjects (HS, mean age of 49.17 ± 11.79; two males and six females) and five psychiatric patients (PP, mean age of 50.25 ± 11.47; two males and three females) with schizophrenia or bipolar disorder. As these are the patients with HD that are often treated with psychotropic drugs, we included PP with similar treatment profiles as a control group in order to minimize the possible iatrogenic impact on the profile of the small non-coding RNAs. In particular, we selected patients treated with olanzapine, lithium, and valproate which were the treatments most frequently prescribed to patients affected by HD. The microarray results indicated that SNORD13 levels were increased in the plasma of patients with HD compared to those in the HS and PP control groups (fold change, 1.54; p = 0.0003 HD vs. HS, and fold change, 1.44; p = 0.0026 HD vs. PP; Figure 1A). To validate this result, SNORD13 plasma levels were quantified using real-time PCR in five cohorts of subjects: 22 HS, 23 symptomatic patients with HD, 15 patients with pre-manifest HD (pre-HD), 24 PP, and 28 patients with Alzheimer’s disease (AD). The last group was considered as a control for the disease specificity of our findings. Demographic and clinical characteristics of each group are shown in Table 1. No significant relationship was observed between SNORD13 plasma levels and age/sex at blood sampling in any group, except for females in the HS group (p = 0.04). No relationship was observed between SNORD13 levels and CAG repeat length in the subjects with pre-HD or HD (Table 2). Our analysis showed a statistically significant (p < 0.0001) increase in the plasma levels of SNORD13 in patients with HD, which clearly segregated patients with overt disease (HD) from controls and pre-HD subjects. The changes in the plasma level of SNORD13 in symptomatic HD patients were highly significant compared to those of both the pre-HD and the three control groups (HS, PP, and AD; Figure 1B, Table 2). A positive linear correlation was observed between circulating SNORD13 levels and disease duration in patients with HD (r = 0.589, p = 0.003) (Figure 2A). A significant relationship was also observed between plasma SNORD13 and the UHDRS clinical score in mHTT carriers (r = 0.732, p < 0.001, Figure 2B), whereas it was not found to be significant in only overt patients with HD (not shown). These linear relationships remained significant when considering age as a covariate (partial correlations, r = 0.563, p = 0.01, and r = 0.685, p < 0.001, respectively). Next, we assessed the accuracy of plasma SNORD13 as a biomarker of overt HD through ROC curve analysis. In discriminating symptomatic HD patients from pre-symptomatic HTT mutation carriers, SNORD13 displayed an extremely high accuracy (AUC = 0.963, Figure 3A); setting the cut-off point of SNORD13 levels at 0.58 allowed the identification of patients with HD with 95.88% sensitivity and 86.7% specificity. Moreover, SNORD13 appeared to be of potential utility in distinguishing symptomatic HD patients from control groups (AD, PP, HS; AUC = 0.953; Figure 3B), as well as pre-HD among controls (AUC = 0.955; Figure 3C), and to a lesser extent, in identifying HTT mutation carriers (AUC: 0.811; Figure 3D). Finally, to investigate the biological landscape of the action of SNORD13, we constructed an interactome. We retrieved information on U13 snoRNA interactions with proteins (including known RNA-binding proteins (RBPs) and transcription factors (TF)), other snoRNAs, miRNAs, and rRNAs from the databases RNAinter [16], snoDB [17], and RNAct [18]. Additionally, we mapped the intra-network protein-protein interactions through STRING [19]. The final network comprised 456 SNORD3-interacting nodes: 258 TFs, 86 RBPs, 91 proteins, one long noncoding RNA, three miRNAs, 13 snoRNAs, and the 18 s rRNA ribosomal subunit (Figure 4A and Supplementary Table S1). A pathway analysis revealed enrichment of processes involved in transcriptional regulation and RNA metabolism (Figure 4B–D), referring to molecules mostly located in the nucleus and involved in genomic organization (Figure 4E). Of interest in HD is the emergence of nerve growth factor (NGF)-stimulated transcription associated with SNORD13 activity, suggesting a direct implication in neurodegenerative processes, as well as the interaction with molecules involved in the DNA damage response that has already been implicated in disease pathogenesis and that are useful as peripheral biomarkers [6,7,20].
Our study highlights an unprecedented finding of the potential role of snoRNAs in HD. The main results are the following: (i) the levels of circulating SNORD13 are significantly increased in the overt disease compared to the prodromal phase of HD; (ii) the levels of this snoRNA are comparable between HS and patients with a pre-HD status; (iii) this finding seems specific for HD, since three groups of control (HS, PP on drugs similar to that administered to patients with HD, and patients with AD) showed normal, comparable values; (iv) the plasma levels of SNORD13 seems to be related to the natural history of HD, correlating with the status of mHTT carriers and the disease duration; and (v) the above points (iii) and (iv) suggest that increased levels of SNORD13 in blood mirror pathogenic events in the CNS (though this fact awaits formal demonstration), possibly paving the way for new therapeutic targets. A possible explanation for our data is that the elevated plasma level of SNORD13 in symptomatic HD patients may be due to nucleolar stress caused by the presence of mutant RNAs that carry an expanded CAG repeat (expanded CAG RNAs). In recent years, increasing attention has been directed toward understanding the pathogenic mechanisms of mHTT, and several studies have supported the hypothesis that expanded CAG RNAs induce apoptosis by activating the nucleolar stress pathway in both patients with HD and transgenic models of the disease [21,22]. Specifically, it has been shown that expanded CAG RNAs compete with nucleolin (a multifunctional protein that is mainly localized in the nucleolus and involved in various steps of ribosome biogenesis) for the rRNA promoter, leading to the reduction of rRNA expression, nucleolar stress and apoptosis via p53, and activation of the downstream signaling cascade, including mitochondrial cytochrome C release and caspase activation [23,24,25,26]. However, considering the SNORD13 interactome, we cannot rule out the possibility that other pathogenic mechanisms might be involved in the change in the levels of snoRNAs. Indeed, beyond ribosomal biogenesis and RNA metabolism, SNORD13 appears to be involved in a wide range of genomic activities associated with HD pathophysiology (Figure 4, Supplementary Table S1): chromatin remodeling via histone acetylation and methylation (SIRT6, EP300, TAF1, CHD1, KDM5A-B etc.), telomere length maintenance (DKC1,HRNPU, VPRBP) [5], DNA repair and damage response (ERCC6, BAZ1B, FANCD2, TOPBP1) [6,20]; direct modulation of homeobox (HOXA1-7, DLX1-2, OTX2 and others), and zinc-finger proteins (SNAI2, SP1-2, ZMIZ2 and others) [27]. We also identified three miRNAs (hsa-miR-455-5p, hsa-miR-342-3p, and hsa-miR-377-3p) which may cooperate with U13 snoRNA in regulating gene expression as being possibly affected in HD; stochastic computational analyses might help further explain these relationships, as piloted in multiple sclerosis pathogenesis modeling [28], COVID epidemic wave stability evaluation [29], and many other complex problems. Irrespective of the mechanism(s) possibly linking mHTT to SNORD13 and the source of this snoRNA (currently unknown), we can speculate that the increase in the plasma level of SNORD13 in patients with HD may peripherally report a ‘tipping point’ in the pathogenic cascade at the neuronal level, while normal levels may mark a ‘molecular pre-manifest status’ in disease evolution. In fact, when we used published data [2] on CSF mHTT and plasma or CSF NfL as benchmarks to compare the performance of plasma SNORD13, we found that it outperformed both CSF mHTT and plasma or intrathecal NfL as a reporter of overt HD, supporting its potential value as a peripheral cue of central pathogenic processes (Table 3) In a clinical context circulating SNORD13 does not have a predictive value, being absent in pre-HD; however, this easily measurable, inexpensive and reliably quantifiable test (it is consistent, accurate, sensitive, specific, and reproducible) [9] may be useful to better catch, on a laboratory basis, the shift from prodromal to overt HD [30]. This may be useful to stratify and select candidate patients for clinical trials and to aim at disease-specific pathophysiological cascades. In summary, circulating SNORD13 could be a clinically actionable asset in future trials as a peripheral reporter of a ‘functional reserve’ at the neuronal level, as well as a hint for new therapeutic targets. Our research is subject to several limitations: the first is the small sample size, but we performed a sample size calculation to reduce this type of bias; the second limitation is the selection bias due to the different age of onset of each disease group taken into exam (AD patients are usually older than HD patients while PP are usually younger); and lastly, our results derive from a cross-sectional study and they consequentially reflect SNORD13 plasma levels at a single point in time both in HD, pre-HD and controls. Further longitudinal studies with a larger sample size will be necessary to reduce these limitations and to validate our hypothesis. In conclusion, peripheral small RNAs may offer potential advantages in terms of disease specificity compared to other approaches, such as circulating NfL, which has already been reported as a promising peripheral biomarker for many neurological diseases [3]. Future personalized therapies for different phases of the HD course and possible trials with etiologic approaches in subjects with pre-HD are necessary, which will plausibly introduce circulating biomarkers such as SNORD13 as promising components of the neurogeneticist’s toolkit.
Participants in the study were enrolled at the Center for Experimental Neurological Therapies, Unit of Neurology (patients with a positive test for HD, patients with probable Alzheimer’s disease, and healthy subjects), and the School of Medicine and Psychology, Unit of Psychiatry (patients with psychiatric disorders), S. Andrea Hospital, Department of Neurosciences, Mental Health, and Sensory Organs, Sapienza University of Rome, Italy. The study was approved by the local ethics committee and written consent was obtained from all participants according to the principles of the Declaration of Helsinki. The operators were unaware of the disease state of each sample during processing and statistical analysis. The eligible subjects for this study were patients with a positive genetic test for HD, a diagnosis of schizophrenia or bipolar disorders treated with antipsychotic drugs, a diagnosis of probable Alzheimer’s disease, and healthy subjects (controls). Exclusion criteria were pregnancy, breastfeeding, and severe systemic illnesses or conditions.
Blood samples were obtained via venous puncture in ethylenediaminetetraacetic acid (EDTA) tubes for plasma preparation. The plasma was obtained by centrifugation (1500× g for 15 min at 4 °C); a few 500 μL aliquots of supernatant were stored at −80 °C. RNA was extracted using a Plasma/Serum Circulating RNA Purification Kit (NORGEN) following the manufacturer’s instructions. RNA quality and purity were assessed using an RNA 6000 Nano Assay kit on an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Briefly, 500 ng of total RNA was labeled using FlashTag Biotin HSR (Genisphere, Hatfield, PA, USA) and hybridized to GeneChipR miRNA 2.0 Arrays. The arrays were stained in Fluidics Station 450 and scanned using a GeneChip R Scanner 3000 (Affymetrix, Santa Clara, CA, USA). A statistical analysis was performed using Transcriptome Analysis Console (TAC) software (Thermo Fisher Scientific, Waltham, MA, USA). To survey the presence of outliers that could impact the dataset, principal component analysis (implemented in R) was performed to identify possible outliers that needed to be excluded. MiRNA probe outliers were defined according to the manufacturer’s instructions (Affymetrix), and further analysis included data summarization, normalization, and quality control using the web-based miRNA QC Tool software (Affymetrix). The raw dataset is available from the Gene Expression Omnibus (GEO) repository (GSE167630).
RNA extraction was performed from plasma using the miRNeasy Serum/Plasma kit according to the manufacturer’s protocol (Qiagen, Hilden, Germany). A U13 snoRNA analysis was performed by quantitative reverse transcription PCR (qRTPCR). cDNA was synthesized using the TaqMan™ MicroRNA Reverse Transcription Kit (Life Technologies, Carlsbad, CA, USA), according to the manufacturer’s instructions using a U13 snoRNA-specific primer (manufacturer-provided). Quantitative real-time PCR was performed on an ABI 7300 Real-time PCR System (using custom TaqMan® Small RNA Assay (Life Technologies), according to the manufacturer’s instructions. Relative quantitation of U13 snoRNA was performed by the delta Ct method, using U6 snoRNA as an endogenous control [32]. Replicate assays of the same sample were performed to calculate inter-assay variation. The average standard deviation (SD) calculated by measuring the plasma SNORD13 levels of a sample repeated over three different assays was 0.035%. Thus, assuming a normal distribution, samples differing in average SNORD13 levels by as little as 0.069% (1.96 × SD) should be distinguishable by this method at a 95% confidence interval [33].
Statistical analyses were performed using Partek Genomic Suite software (miRNA Array data), GraphPad Prism v9, and R (The R Project for Statistical Computing) v3.6.3. Data normality was assessed using the Shapiro–Wilk test. Nonparametric tests were used to compare the distribution of SNORD13 plasma levels between the patients and controls. Nonparametric tests or linear regression were used to evaluate the distribution of SNORD13 plasma levels across age, sex, and CAG repeat number. Spearman’s correlation was computed to assess the linear relationship between SNORD13 plasma levels and disease duration or the Unified Huntington’s Disease Rating Scale-Total Motor Score (UHDRS-TMS), and age was considered a covariate in the partial correlations. Statistical significance was set at p < 0.05. The sample size calculation was based on our preliminary study that enrolled 10 HD subjects and 10 controls, detecting a difference between group means of 0.58 with a standard deviation of 0.34. Assuming α = 0.01 and equal group size, we would need to study 12 HD and 12 control subjects to reject the null hypothesis that the population means of the experimental and control groups are equal with probability (power) 0.9. A receiver operating characteristic (ROC) curve analysis was performed using the R easy ROC web interface. The optimal cutoff point was identified according to the Youden index method. Sample size estimation settings were as follows: type I error, 0.05; power, 0.8; allocation ratio, 1; area under the curve (AUC) derived from the HD/pre-HD ROC curve.
The molecular interactions of SNORD13 were screened in three databases: (1) the RNA interactome database (RNAinter v4.0) [34] at http://www.rna-society.org/rnainter/ (accessed on 1 November 2021), which includes more than 41 million predicted or experimentally validated interactions; (2) snoDB, the largest repository of snoRNA biological annotation and manually curated snoRNA-RNA interactions [35]; (3) RNAct, a database of genome-wide predicted protein–RNAs interactions [17]. Nonhuman interactors were excluded from this study. Protein-protein interactions were derived by querying the STRING database [18] for high-confidence interactions (score > 0.7) of SNORD13-interacting proteins. The global SNORD13 interactome was mapped using Cytoscape [19]. A pathway enrichment analysis was performed using the STRING application in Cytoscape. Graphical plotting was performed using the ggplot2 package in R [20]. | true | true | true |
|
PMC9604388 | Qingtai Wang,Kui Fang,Lizhong Qi,Xiao Wang,Yu Pan,Yunshuo Li,Jinghui Xi,Juhong Zhang | Purification and Functional Characterization of a Soluble Trehalase in Lissorhoptrus oryzophilus (Coleoptera: Curculionidae) | 24-09-2022 | Lissorhoptrus oryzophilus,soluble trehalase,molecular docking,RNAi,prokaryotic expression,homology modeling | Simple Summary The rice water weevil, Lissorhoptrus oryzophilus Kuschel (Coleoptera: Curculionidae), is indigenous to the United States and has become a significant invasive agricultural pest in China. In this study, we identified and cloned one trehalase gene (LoTRE1) encoding a soluble protein in L. oryzophilus and compared the relative expression levels of LoTRE1 in different tissues. The purified LoTRE1 protein was obtained using a prokaryotic expression system, and its enzymatic properties were explored. Amino acid sequence homology modeling of LoTRE1 and molecular docking between the LoTRE1 protein and substrate trehalose were simulated, which further provided a theoretical basis for revealing the role of LoTRE1 in the degradation mechanism of trehalose. In addition, the LoTRE1 double-stranded RNA (dsRNA) was synthesized in vitro, and its RNAi effect in L. oryzophilus was detected via feeding. The results suggested that LoTRE1 played a vital role in L. oryzophilus development, which could be useful for providing information for insect pest control in the future. Abstract Trehalase is the only enzyme known for the irreversible splitting of trehalose and plays a major role in insect growth and development. In this report, we describe a basic study of the trehalase gene fragment encoding a soluble trehalase from Lissorhoptrus oryzophilus (LoTRE1). Sequence alignment and phylogenetic analysis suggested that LoTRE1 was similar to some known insect trehalases and belongs to the Coleoptera trehalase group. Additionally, LoTRE1 was expressed mainly in the fat body. Purified protein was obtained using heterologous expression of LoTRE1 in Escherichia coli, and the recombinant protein exhibited the ability to decompose trehalose. Enzyme–substrate docking indicated the potential involvement of other residues in the catalytic activity, in addition to Asp 333. Moreover, feeding of adults on LoTRE1 dsRNA silenced the transcription of LoTRE1 and thereby reduced the activity of trehalase and increased the trehalose content; it also led to a 12% death rate. This study reveals essential molecular features of trehalase and offers insights into the structural aspects of this enzyme, which might be related to its function. Taken together, the findings demonstrate that LoTRE1 is indispensable for adults of this pest and provide a new target for the control of L. oryzophilus. | Purification and Functional Characterization of a Soluble Trehalase in Lissorhoptrus oryzophilus (Coleoptera: Curculionidae)
The rice water weevil, Lissorhoptrus oryzophilus Kuschel (Coleoptera: Curculionidae), is indigenous to the United States and has become a significant invasive agricultural pest in China. In this study, we identified and cloned one trehalase gene (LoTRE1) encoding a soluble protein in L. oryzophilus and compared the relative expression levels of LoTRE1 in different tissues. The purified LoTRE1 protein was obtained using a prokaryotic expression system, and its enzymatic properties were explored. Amino acid sequence homology modeling of LoTRE1 and molecular docking between the LoTRE1 protein and substrate trehalose were simulated, which further provided a theoretical basis for revealing the role of LoTRE1 in the degradation mechanism of trehalose. In addition, the LoTRE1 double-stranded RNA (dsRNA) was synthesized in vitro, and its RNAi effect in L. oryzophilus was detected via feeding. The results suggested that LoTRE1 played a vital role in L. oryzophilus development, which could be useful for providing information for insect pest control in the future.
Trehalase is the only enzyme known for the irreversible splitting of trehalose and plays a major role in insect growth and development. In this report, we describe a basic study of the trehalase gene fragment encoding a soluble trehalase from Lissorhoptrus oryzophilus (LoTRE1). Sequence alignment and phylogenetic analysis suggested that LoTRE1 was similar to some known insect trehalases and belongs to the Coleoptera trehalase group. Additionally, LoTRE1 was expressed mainly in the fat body. Purified protein was obtained using heterologous expression of LoTRE1 in Escherichia coli, and the recombinant protein exhibited the ability to decompose trehalose. Enzyme–substrate docking indicated the potential involvement of other residues in the catalytic activity, in addition to Asp 333. Moreover, feeding of adults on LoTRE1 dsRNA silenced the transcription of LoTRE1 and thereby reduced the activity of trehalase and increased the trehalose content; it also led to a 12% death rate. This study reveals essential molecular features of trehalase and offers insights into the structural aspects of this enzyme, which might be related to its function. Taken together, the findings demonstrate that LoTRE1 is indispensable for adults of this pest and provide a new target for the control of L. oryzophilus.
Trehalose is a nonreducing disaccharide that consists of two α-glycosidically linked glucose units. This disaccharide is found in many organisms, such as plants, nematodes, bacteria, insects, and some other invertebrates; however, it does not exist in mammals [1]. Trehalose is mainly considered an important cell protective metabolite in vivo, and cells can synthesize trehalose in large quantities under stress and degrade it rapidly under normal conditions [2]. Trehalose exhibits multiple physiological effects in various organisms, and it has been shown that trehalose can shield proteins and cellular membranes from inactivation or denaturation caused by a range of stress conditions, including dehydration, desiccation, cold, heat, and oxidation [3,4,5,6,7,8]. Additionally, trehalose is the major blood sugar in insects, playing an important role as an instant source of energy and in the response to abiotic stresses. Trehalose is specifically synthesized in insect fat bodies and quickly discharged into the hemolymph and other tissues [7,8]. To utilize hemolymph trehalose, insect tissues contain trehalases (EC 3.2.1.28) that catalyze the hydrolysis of one mole of trehalose to two moles of glucose. Thus, trehalase is the enzyme that is required for the uptake or utilization of trehalose in the hemolymph of insects. Two types of trehalase, soluble trehalase (TRE1) and membrane-bound trehalase (TRE2), have been cloned and characterized in several insect species, such as Acyrthosiphon pisum, Bombyx mori, Helicoverpa armigera, and Harmonia axyridis [9,10,11,12]. Both TRE1 and TRE2 include a signal peptide, two signature motifs (PGGRFREFYYWDSY and QWDYPNAWPP), and one glycine-rich region [13]. Gibson et al. [14] have shown that the catalytic domain of E. coli trehalase displays an aspartate (Asp312) and a glutamate (Glu496) residue, which play the role of a general acid and a general base, respectively, similarly to hydrolases from the GH37 family. By site-directed mutagenesis in Spodoptera frugiperda, three arginine residues essential to the enzyme activity were identified inside the active site [15]. To date, no three-dimensional (3D) structure is available from experimental data for TREs from plants, animals, or fungi, whereas molecular modeling studies have predicted the 3D structure of insect TREs in Bombyx mori [10], Helicoverpa armigera [11], Drosophila melanogaster [16], Chironomus riparius [17], and Delia antiqua [18]. There are also individual protein differences among different insects [10,17]. In vivo, soluble trehalase accounts for the majority of the overall trehalase enzyme activity, according to previous studies [19]. The main function of TRE1 is to decompose trehalose in cells. TRE1 is an essential enzyme in insect energy metabolism and the first enzyme in the chitin synthesis pathway in insects [11]. The rice water weevil, Lissorhoptrus oryzophilus Kuschel (Coleoptera: Curculionidae), is the most harmful and widely distributed early-season pest of rice in the USA [20] and causes serious economic issues in wetland rice agriculture, resulting in losses of up to 25% in untreated fields [21]. Parthenogenetic female Lissorhoptrus oryzophilus (L. oryzophilus) have quickly invaded many rice-growing regions around the world [20,22,23]. In China, L. oryzophilus were first discovered in 1998; they have rapidly invaded 78% of the provinces and have become the foremost widespread invasive pest [24]. L. oryzophilus have caused serious harm to the rice in the invaded area. In this paper, we analyzed the sequence, structural characteristics, and expression patterns of LoTRE1. We also expressed recombinant LoTRE1 and studied its physicochemical properties to gain insight into the optimum conditions for its activity, including temperature and pH. Furthermore, we used RNAi to study the function of the LoTRE1 gene and detected changes in target gene expression, trehalase activity, and trehalose content. The findings of our research are critical for understanding the performance of trehalase at the molecular level, as well as the potential to adapt to future invasions. The characterization of trehalase genes could facilitate the development of novel ways to manage L. oryzophilus.
L. oryzophilus adults were collected from Changchun, Jilin Province, China, and reared on rice seedlings under a 16 h light/8 h dark photoperiod at 26 ± 1 °C with 80 ± 5% relative humidity. The adults and different tissues (hemolymph, midgut, head, wing, fat body, and leg) of dissected L. oryzophilus were promptly immersed in liquid nitrogen and stored at −80 °C until use.
Total RNA was extracted from adults and different tissues using RNAiso Plus (Takara, Dalian, China) according to the manufacturer’s instructions for gene cloning and spatial expression. The RNA integrity and concentration were checked using agarose gel electrophoresis and spectrophotometry (NanoDrop2000, Wilmington, DE, USA), respectively. Then, approximately 1 μg of total RNA was utilized for the synthesis of first-strand cDNA by using the PrimeScript™ RT Reagent Kit with gDNA Eraser (Takara, Dalian, China). The synthesized cDNA was stored at −20 °C until use.
The sequence of LoTRE1 was identified using our unpublished transcriptome data. The amino acid sequence of LoTRE1 was deduced using DNAMAN software. Open reading frames (ORFs) of genes were predicted using ORF finder (https://www.ncbi.nlm.nih.gov/orffinder/, accessed Date: 25 June 2021). Signal peptides and transmembrane domains were predicted using SignalP-5.0 (http://www.cbs.dtu.dk/services/SignalP/, accessed Date: 25 June 2021) and TMHMM2.0 (http://www.cbs.dtu.dk/services/TMHMM/, accessed Date: 25 June 2021), respectively. The molecular weight and theoretical isoelectric point of the deduced protein were calculated using the ExPASy Compute pI/Mw tool (http://web.expasy.org/compute_pi/, accessed Date: 25 June 2021) [25]. Multiple amino acid sequence alignments were performed by using DNAMAN software (http://www.lynnon.com/pc/alignm.html, accessed Date: 25 June 2021). The tertiary structures of LoTRE1 proteins were predicted using SOPMA (https://npsa-prabi.ibcp.fr/cgi-bin/secpred_sopma.pl, accessed Date: 25 June 2021) and SWISS-MODEL (https://swissmodel.expasy.org/, accessed Date: 25 June 2021). PDB files and molecular ligand data were obtained from ZINC (http://zinc15.docking.org/substances/home/, accessed Date: 25 June 2021). The molecular model docking calculations of LoTRE1 with ligands were performed by using the Autodock 4.2 program. BLASTX best hits were found using the BLASTX program provided by NCBI (http://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed Date: 25 June 2021). Phylogenetic trees were constructed using MEGA 6.0 software with maximum-likelihood phylogenetic analysis. The tree was colored and arranged using iTOL (https://itol.embl.de/upload.cgi, accessed Date: 25 June 2021).
Primer pairs for qPCR were designed using Primer 5 software, as shown in Table 1. GAPDH was used as a reference gene. mRNA expression levels were measured using qPCR using SYBR qPCR SuperMix (TransGen, Beijing, China). Each amplification reaction was carried out in a total volume of 20 μL, with 1 µL of cDNA, 10 µL of green qPCR SuperMix, 0.4 µL of forward primer, 0.4 µL of reverse primer, and 8.2 µL RNase-free water. qPCR was performed on an ABI 7500 Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA) under the following conditions: initial denaturation at 94 °C for 30 s, followed by 40 cycles of denaturation at 94 °C for 5 s, annealing at 55 °C for 15 s, and extension at 72 °C for 10 s. To calculate the relative expression levels, melting curves were evaluated to confirm the single peak and check amplification specificity after qPCR. Standard deviations and means were obtained from the mean of three biological replicates with three corresponding technical replicates. The relative expression value of the LoTRE1 gene was calculated using the 2−ΔΔCt method [26].
The coding region of the LoTRE1 gene was subcloned into the BamHI/XhoI sites of the pET28a (+) vector and then transformed into Rosetta (DE3) E. coli competent cells. The colonies were grown on Luria–Bertani culture medium with kanamycin (50 mg/mL). The positive monoclones were cultured in liquid LB medium (supplemented with 50 mg/mL kanamycin) overnight at 37 °C. The culture was diluted at 1:100 in liquid LB and incubated at 37 °C for 3–4 h until the OD600 reached 0.4–0.6. Isopropyl-β-d-thiogalactoside (IPTG) was added at final concentrations of 0.1, 0.4, 0.8, and 1.0 mmol/L, and then the culture was incubated at different temperatures (28 and 37 °C) for 8 h. Centrifugation (5000 rpm, 5 min, 4 °C) was performed to harvest the cells, and the collected cells were suspended in 1× phosphate-buffered saline (PBS). The suspension was sonicated on ice and centrifuged (5000 rpm, 5 min, 4 °C) for a second time. Protein present in the supernatant was purified with a Ni-NTA Gravity Column (Sangon, Shanghai, China). The purified protein was assessed using SDS–PAGE and then quantified via the Bradford assay with BSA as the standard.
The purified protein was separated using 10% SDS–PAGE and then transferred to a polyvinylidene fluoride (PVDF) membrane (100 V, 1 h). The membrane was treated with 5% blocking protein powder in TBST (1 M Tris-HCl pH 7.5, 500 mM NaCl, and 0.2% Tween 20) for 1 h at room temperature and reacted with the anti-6 × His tag mouse monoclonal antibody (Sangon, Shanghai, China) in TBST for 1.5 h at room temperature. The membrane was washed with TBST 4 times for 5 min each time and then incubated with AP-conjugated goat anti-mouse IgG (Sangon, Shanghai, China) in TBST for 1 h at room temperature. The NBT/BCIP substrate solution (Sangon, Shanghai, China) was utilized to visualize the protein band after the PVDF membrane was washed again.
The 3,5-dinitrosalicylic acid technique was used to indirectly measure trehalase activity [27]. The reaction mixture (1 mL) consisted of 10 µL purified protein, 50 µL trehalose (200 mmol/L), and 940 µL PBS. PBS was prepared for a pH range of 3.0–7.0 The mixture was incubated in each pH buffer at 25 °C for 30 min and subjected to boiling for 5 min. The coagulated protein was removed by centrifugation at 12,000 rpm for 10 min at 4 °C. Trehalase activity was determined by measuring the content of glucose released during incubation. Similarly, for the measurement of trehalase activity at different temperatures, the mixture containing PBS (pH 7.0) and other components was incubated at the 9 individual temperatures, which ranged from 5 to 75 °C. To determine the kinetic parameters (Km and Vmax) of trehalase, substrates at different concentrations (1, 2.5, 5, 7.5, and 10 mmol/L) were added to the reaction mixture and incubated at 50 °C and pH 7.0 for 30 min. The trehalase activity was recorded with a UV-2450 spectrophotometer (Shimadzu, Japan). One unit of enzyme activity (U) was defined as the amount of protein that released 1 μmol of glucose in 1 min. Each experiment was replicated three times.
The dsRNA of LoTRE1 was synthesized using the T7 RioMAX Express RNAi System (Promega, San Luis Obispo, CA, USA) according to the manufacturer’s recommendations. Green fluorescent protein (GFP) dsRNA was used as the control. The primers used to synthesize the dsRNA are listed in Table 1. A total of 270 L. oryzophilus adults were divided into three groups, and each group was fed rice leaves with dsGFP, dsLoTRE1, and RNase-free water. Approximately 30 individuals were fed per treatment. Each group contained three biological replicates. Edges of approximately 1 cm were cut from both ends of fresh rice leaves, and the leaves were dried for 30 s at 55 °C in a drying oven. The rice leaves were immersed in 10 mL centrifuge tubes containing 500 ng/µL dsRNA for 6 h. The adults of L. oryzophilus were transferred to a new 10 mL centrifuge tube, and each tube contained 30 insects. The treated rice leaves were put into the centrifuge tubes containing L. oryzophilus adults, and then the centrifuge tubes were sealed with gauze. dsRNA feeding was continued for 12 h, and the leaves were replaced with fresh rice leaves without dsRNA every 24 h. The dead insects were picked out with a brush, and the number of dead insects was recorded. Fresh leaves and dsGFP-treated leaves were used as controls. Total RNA from treated L. oryzophilus adults was extracted as a template, and specific primers (Table 1) were used for quantitative RT–qPCR. Each RT–qPCR biological replicate contained three surviving L. oryzophilus adults. Three biological replicates and three technical replicates were set for each treatment group.
A total of 30 L. oryzophilus adults were placed in a 10 mL centrifuge tube and fed with rice leaves treated with dsRNA for 12 h, then transferred to fresh rice leaves without dsRNA. The surviving individuals were picked out with a brush after 48 h. The enzymatic analysis of the activity of trehalase (THL) was performed using a THL kit according to the manufacturer’s instructions (Solarbio, Beijing, China), and the trehalose content was determined according to the manufacturer’s instructions for the Trehalose Content Kit (Solarbio, Beijing, China). Each experiment was replicated three times.
All data obtained during this study were expressed as the means ± standard deviations of three replicates and were tested with a one-way ANOVA and t tests using SPSS 22.0. p values of less than 0.01 indicated significant (**) differences.
LoTRE1 was identified from our unpublished transcriptome data of L. oryzophilus. Through homology searching in our transcriptional sets, the sequence of LoTRE1 was found to contain an 1839 bp-long open reading frame encoding 612 amino acid residues (approximately 70833 Da and a theoretical isoelectric point (pI) of 5.38). A signal peptide of 25 amino acids and a cleavage site (IAI-YK) between residues 25 and 26 were identified (Figure 1). LoTRE1 had two signature motifs (PGGRFREFYYWDSY and QWDYPNAWPP) and a highly conserved glycine-rich region (GGGGEY). Four N-glycosylation sites (residues 336, 579, 580, and 581) were identified in the LoTRE1 sequence (Figure 1). TMHMM predicted that there were no transmembrane domains. Multiple sequence alignment was performed to analyze evolutionarily or structurally related positions between LoTRE1 and its homologues based on the amino acid sequence (Figure 2). LoTRE1 showed approximately 55–72% sequence identity with other Coleoptera insect trehalases. The greatest sequence identity (72%) was with trehalase (GenBank ID: XP 030756586.1) from Sitophilus oryzae, and the lowest sequence identity (55%) was with trehalase (GenBank ID: XP_031336857.1) from Photinus pyralis. To investigate the evolutionary relationships of LoTRE1, a phylogenetic tree of LoTRE1 was constructed along with several previously studied TREs from 24 species of Coleoptera, Hemiptera, Blattaria, and Hymenoptera using maximum-likelihood phylogenetic analysis (Figure 3). The results showed that LoTRE1 shared the highest homology with the genes from Sitophilus oryzae and Rhynchophorus ferrugineus.
The expression profiles of LoTRE1 in numerous tissues were probed using qPCR. LoTRE1 was shown to be expressed in a variety of tissues, including the head, midgut, hemolymph, wing, leg, and fat body (Figure 4). Remarkably, LoTRE1 had high expression within the fat body; however, it had the lowest expression in the wing. The relative expression in the fat body was 120 times that in the wing.
The SWISS-MODEL service was used to predict the three-dimensional structure of LoTRE1 based on the resolved crystal structure of the periplasmic trehalase of E. coli (PDB ID 2WYN; Figure 5). A Ramachandran plot and the Procheck server were used to verify the homology model’s dependability (Figure 6). A total of 89.7% of amino acid residues were in the most favored regions, 9.9% of amino acid residues were in the additional regions, and 0.2% of amino acid residues were in the disallowed regions. Thus, this was a high-quality model. The predicted α-toroidal structure of LoTRE1 was very similar to that of the periplasmic trehalase of E. coli, but one β-sheet was absent in LoTRE1 at position 1. A molecular docking study was performed to further understand the interaction of trehalose with trehalase and to obtain insight into the binding location of this molecule on trehalase. The 3D protein structure with small molecular compounds was studied using molecular docking simulations (Figure 7). The results showed that trehalose bound tightly to LoTRE1 (G = −1.36 kcal/mol). The docking results revealed that the interactions of trehalose with LoTRE1 were mediated by ARG 179, TRP 186, ASN 223, ARG 232, GLN 234, ARG 297, GLU 299, GLY331, ASP 333, and TRP 477.
To obtain recombinant LoTRE1 protein in large quantities, we induced protein production in E. coli Rosetta (DE3) cells under different induction conditions (Figure 8A). The molecular weight of the trehalase protein was found to be close to the predicted size in all the tests (70 kDa). Therefore, we chose a final concentration of 0.1 mmol/L IPTG at 28 °C to obtain large quantities of recombinant protein. In addition, the LoTRE1 recombinant protein was mostly expressed as soluble protein (Figure 8B). The purified LoTRE1 protein was analyzed using SDS–PAGE (Figure 8C), and one obvious unique band appeared at the predicted size (70 kDa). The expression of the recombinant LoTRE1 protein was verified using Western blotting (Figure 8D). The concentration of the purified LoTRE1 protein was 0.98 mg/mL.
The enzyme activity increased with increasing temperature and pH. The optimum activity of LoTRE1 was found at 50 °C and pH 7.0, beyond which the enzyme activity decreased. No catalytic ability was found at temperatures above 75 °C or below 5 °C, indicating that the catalytic performance of LoTRE1 had a restricted temperature range. Under the optimal conditions, the specific activity of purified LoTRE1 was 58.47 ± 1.74 U/mg. The kinetic parameters km and Vmax of LoTRE1 were 48.6 mmol/L and 1.108 mmol/(L·min), respectively (Figure 9).
qPCR was performed to prove the effectiveness of RNAi in L. oryzophilus. Feeding dsRNA-LoTRE1 significantly reduced the expression levels of trehalase in the rice water weevil compared with that in the nontarget control group within 24 and 48 h (p < 0.01). The results showed that RNAi reduced the expression levels of LoTRE1 to 40 % at 24 h and 56 % at 48 h in L. oryzophilus (Figure 10A). The mortality rate of L. oryzophilus increased significantly (p < 0.01) at 24 h (9%) and 48 h (12%) after dsTRE1 feeding (Figure 10B).
At 48 h after feeding dsRNA-LoTRE1, the trehalase activity of L. oryzophilus was 13.14 ± 0.47 U/g, which was 31.6% lower than that of the control group (Figure 11A), and there was a significant difference compared with the control group (p < 0.01). The trehalose content of L. oryzophilus was 5.73 ± 0.60 mg/g, which was 130% higher than that of the control group (Figure 11B), and there was a significant difference compared with the control group (p < 0.01).
Trehalase catalyzes the hydrolysis of trehalose and plays a vital role in insect metabolism. The amount of trehalases has been found to differ among insect species. A total of three trehalase genes were found in N. lugens, and three trehalase genes were identified in Tribolium castaneum [28,29]. One trehalase gene, LoTRE1, was identified from our previous transcriptome database of L. oryzophilus, and the analysis of the deduced amino acid sequence demonstrated that LoTRE1 encodes a soluble trehalase. LoTRE1 contained some conserved regions, including one signal peptide, two signature motifs, four putative glycosylation sites, and a highly conserved glycine-rich sequence. The findings were consistent with those for trehalase genes found in other Coleoptera species. Amino acid sequence alignment showed that LoTRE1 shared high identity with some known insect trehalases. The LoTRE1 sequence shared the highest identity with that of S. oryzae (72% identity GenBank ID: XP 030756586.1), followed by those of R. ferrugineus (70% identity GenBank ID: KAF7271780.1), A. planipennis (65% identity GenBank ID: XP_018322659.1), H. axyridis (59% identity GenBank ID: AOT82130.1), and P. pyralis (55% identity GenBank ID: XP_031336857.1). The phylogenetic tree revealed that LoTRE1 shared the highest homology with the genes from Sitophilus oryzae and Rhynchophorus ferrugineus. The model of LoTRE1 shared a general α-toroidal architecture with the template protein. According to the CAZy database, insect trehalases are mostly members of GH family 37 and possess glutamic acid (Glu) as a nucleophile, while their proton donor is aspartic acid (Asp) [16]. We identified Asp 333 and Glu 299 as important catalytic residues in the LoTRE1 model by comparing it to the E. coli periplasmic trehalase. The trehalose molecule occupies a space in the active center pocket of the enzyme that contains potential acid (Asp 333) and base (Glu 299) residues, as well as three conserved Arg residues (R 179, R 232, R 297). Our results with LoTRE1 were consistent with those reported in other insect trehalases [9,30]. The primary three-dimensional structure of LoTRE1 was similar to the periplasmic trehalase of E. coli. Both structures mainly consisted of α-helixes and were surrounded by α-toroidal structures. This structure was also observed in the trehalase of other insects [9,17]. According to Silva et al., guanidine groups and Arg residues are required for soluble trehalase activity, and chemical alteration of these residues results in enzyme inactivation. This was later confirmed by site-directed alterations in three Arg residues within the enzyme’s active site pocket [15]. In some insect species, TRE1 has been purified from the goblet cell cavity, hemolymph, egg homogenates, and midgut [31]. The tissue expression pattern analysis revealed that LoTRE1 was most highly expressed in the fat body, followed by the midgut. Yu et al. (2021) found that trehalase was most highly expressed in the head and wings of Diaphorina citr [32]. Ma et al. (2015) found the highest expression of trehalase in the midgut in Helicoverpa armigera [33]. These results indicate that TRE1 could serve completely different functions in several tissues during the development of various insects. The insect fat body functions as an energy storage center. It is also the organization center of the metabolic processes of insects, such as growth, development, metamorphosis, and reproduction. Trehalase could cooperate with the hormones in fat to dynamically control the concentrations of trehalose and glucose in insects [34]. Trehalases can be divided into acidic trehalases and neutral trehalases according to their optimal pH. Inagaki et al. (2001) found that the optimum pH of the trehalase from Acidobacterium capsulatum was 2.5 [35]. Lee et al. (2001) revealed that the optimum pH of the trehalase from Apis mellifera was 6.7 [36]. In this study, the recombinant LoTRE1 protein had enzyme activity in the range of pH 3.0~9.0, but the highest activity was observed at pH 7.0, which indicated that the LoTRE1 protein has the best enzyme activity in neutral environments. Previous studies have shown that the optimum temperature of trehalase is generally high, usually at 40~65 °C [37]. In this experiment, LoTRE1 had no activity at temperatures below 5 °C, and the enzyme activity increased gradually with increasing temperature. When the system temperature exceeded 75 °C, LoTRE1 activity was lost. Shukla et al. (2016) revealed that the optimum temperature of the trehalase from Drosophila melanogaster was 55 °C [16], Ai et al. (2018) revealed that the optimum temperature of the trehalase from Helicoverpa armigera was 55 °C [11]. Since trehalase remains the only enzyme known for the irreversible splitting of trehalose under physiological conditions, heterologous expression of LoTRE1 can be used to explore the biochemical characteristics and kinetic properties of the enzyme. RNAi is a highly conserved mechanism initiated by sequence-specific double-stranded RNA (dsRNA), leading to target-specific endogenous gene silencing [38]. RNAi has been widely accepted as a powerful tool for gene function research and is relatively well-established in a variety of insects, such as Bemisia tabacii [39] and Hemiptera [40]. Previous studies have shown that silencing trehalase in Leptinotarsa decemlineata caused larval death [32]. Silencing of trehalase in Nilaparvata lugens caused phenotypic deformities [41]. These results suggested that trehalase has a biological function in insect development and survival. In this study, RNAi was performed to determine whether the levels of LoTRE1 would affect L. oryzophilus development and survival. The results indicated that dsRNA feeding-mediated silencing of the LoTRE1 gene caused not only downregulation of the transcript level of LoTRE1 but also decreased trehalase activity in treated adults. In contrast, silencing of LoTRE1 increased the trehalose content and resulted in lethal effects in adults. In agreement with our results, knockdown of the trehalase gene led to mortality in treated Harmonia axyridis and Spodoptera exigua [12,42]. In addition, silencing HaTRE1 led to significant decreases in the ability of females to attract males and successful mating proportions [43]. Trehalose has been well demonstrated in insect physiology as an energy source for insects, maintaining the glucose level [44]. The energy level and blood glucose content are affected in various cells of insects when the hydrolysis of trehalose to glucose is inhibited, and thus other physiological pathways are affected [42,45]. Trehalase can decompose the important energy storage material and the stress metabolite trehalose in insects. The changes in gene expression and enzyme activity affect the life processes of insects, including molting, metamorphosis, and reproduction [34]. Studies have shown that the TRE gene affects the levels of three sugars by regulating gene expression and enzyme activity [44,46]. These results collectively indicated that the LoTRE1 gene plays a vital role in L. oryzophilus by regulating the trehalose content. Feeding with dsTRE1 can disrupt the metabolism of trehalose in the body. The results lay a foundation for exploring the potential functions and regulatory mechanisms of insect TRE1, which could be useful for providing information for insect pest control in the future. In conclusion, we identified one soluble trehalase gene from L. oryzophilus, analyzed its molecular characteristics, and explored its biochemical characteristics, kinetic properties, and optimum reaction conditions through heterologous expression analysis. The importance of LoTRE1 was inferred through RNAi. LoTRE1 is a key gene regulating the expression of trehalase and the trehalose content. However, the particular role of this gene is still unknown, and more research is needed to gain better knowledge of LoTRE1’s functions. The results of this study provide a good basis for further studies on the regulation of the expression of this gene and provide a potential target for the control of L. oryzophilus in the field. | true | true | true |
Subsets and Splits